ISBN 974-229-997-2
ราคา 220 บาท
Strengthening The Hard Disk Industry in Thailand National Electornics and Computer Technology Center National Science and Technology Development Agency Ministry of Science and Technology 112 Thailand Science Park, Phahon Yothin Road, Klong Luang, Pathumthani 12120, THAILAND. Tel. +66(0)2-564-6900 Fax. +66(0)2-564-6901..2
การประมวลผลสัญญาณสำหรับการจัดเก็บข้อมูลดิจทิ ลั เล่ม 2 : การออกแบบวงจรภาครับ
โครงการเสริมสร้างความแข็งแกร่งให้กบั อุตสาหกรรมฮาร์ดดิสก์ไดรฟ์ในประเทศไทย ศูนย์เทคโนโลยีอเิ ล็กทรอนิกส์และคอมพิวเตอร์แห่งชาติ สำนักงานพัฒนาวิทยาศาสตร์และเทคโนโลยีแห่งชาติ กระทรวงวิทยาศาสตร์และเทคโนโลยี 112 อุทยานวิทยาศาสตร์ประเทศไทย ถนนพหลโยธิน ตำบลคลองหนึง่ อำเภอคลองหลวง จังหวัดปทุมธานี 12120 โทรศัพท์ 02-564-6900 โทรสาร 02-564-6901..2
ดร. ปิยะ โควินท์ทวีวฒ ั น์
ÊÓËÃѺ¤Ãͺ¤ÃÑÇ â¤ÇÔ¹·ì·ÇÕÇѲ¹ì
¤Ó¹ÔÂÁ
¹Ñº ¨Ò¡» ¾.È. 2548 ÈÙ¹Âì à·¤â¹âÅÂÕ ÍÔàÅç¡·Ã͹ԡÊì áÅФÍÁ¾ÔÇàµÍÃì áËè§ªÒµÔ (๤෤) Êӹѡ§Ò¹ ¾Ñ²¹ÒÇÔ·ÂÒÈÒʵÃì áÅÐà·¤â¹âÅÂÕ áËè§ªÒµÔ (ÊÇ·ª.)
ä´é ÃѺ ÁͺËÁÒÂãËé à» ¹ à¨éÒÀÒ¾´Óà¹Ô¹§Ò¹ªØ´
â¤Ã§¡ÒÃàÊÃÔÁÊÃéÒ§¤ÇÒÁá¢ç§á¡Ãè§ãËéá¡èÍصÊÒË¡ÃÃÁÎÒÃì´´ÔÊ¡ìä´Ã¿ìã¹»ÃÐà·Èä·Ââ´ÂÁÕÇѵ¶Ø»ÃÐʧ¤ì ËÅÑ¡ ¤×Í ¡ÒáÃдѺ ¤ÇÒÁÊÒÁÒö·Ò§¡ÒÃá¢è§¢Ñ¹ ÃÐËÇèÒ§»ÃÐà·È áÅÐÊè§àÊÃÔÁ ãËé »ÃÐà·Èä·Â¤§ ¤ÇÒÁà» ¹¼Ùé¼ÅÔµÎÒÃì´´ÔÊ¡ìä´Ã¿ìáÅÐÊèǹ»ÃСͺ Íѹ´Ñº 1 ¢Í§âÅ¡µÅÍ´·ÈÇÃÃÉ˹éÒ à¾× ÍãËéºÃÃÅØ Çѵ¶Ø »ÃÐʧ¤ì ´Ñ§¡ÅèÒÇ à¹¤ à· ¤ ä´é ´Óà¹Ô¹§Ò¹ ¼èÒ¹ ¡Ô¨¡ÃÃÁ ÃٻẺ µèÒ§æ àªè¹ ¡Òà ʹѺʹع ·Ø¹ÇԨѠà¾× Í ¾Ñ²¹Òà·¤â¹âÅÂÕ ¡ÒüÅÔµ ã¹ÍصÊÒË¡ÃÃÁÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ¡ÒÃʹѺʹع ·Ø¹ à¾× Í ¾Ñ²¹ÒºØ¤ÅÒ¡Ã ·Ñ §¡è͹ ËÅѧà¢éÒ·Ó§Ò¹ã¹ÍصÊÒË¡ÃÃÁÎÒÃì´´ÔÊ¡ìä´Ã¿ì ¡ÒþѲ¹ÒËèǧâ«èÍØ»·Ò¹ ÃÇÁ¶Ö§¡ÒþѲ¹Ò ¹âºÒÂ
ÁҵáÒÃ
áÅÐâ¤Ã§ÊÃéÒ§¾× ¹°Ò¹·Õ à¡Õ ÂÇ¢éͧ¡ÑºÍصÊÒË¡ÃÃÁÎÒÃì´´ÔÊ¡ìä´Ã¿ì
ÍÂèÒ§äáçµÒÁ
๤෤µÃÐ˹ѡ ´Õ ÇèÒ ¹Í¡à˹×Í ¨Ò¡¡Ô¨¡ÃÃÁµèÒ§æ àËÅèÒ¹Õ áÅéÇ ¡Ô¨¡ÃÃÁ·Õ ÊÓ¤Ñ ÍÕ¡ ÃÙ»áººË¹Ö § ¤×Í ¡ÒöèÒ·ʹͧ¤ì ¤ÇÒÁÃÙé áÅÐà·¤â¹âÅÂÕ µèÒ§æ ·Õ à¡Õ ÂÇ¢éͧ¡Ñº ÍصÊÒË¡ÃÃÁÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì à¾ÃÒСÒà ´Óà¹Ô¹§Ò¹ã´æ ÂèÍÁ¡èÍ ãËé à¡Ô´ ͧ¤ì ¤ÇÒÁÃÙé «Ö § ͧ¤ì ¤ÇÒÁÃÙé µèÒ§æ àËÅèÒ¹Õ ËÒ¡äÁè ä´é ¶Ù¡ ÃǺÃÇÁËÃ×Í ¨Ñ´ à¡çºÍÂèҧ໠¹Ãкº áÅÐäÁèä´éà¼Âá¾ÃèÍÍ¡ä» ÂèÍÁà·èҡѺÇèÒ ¡ÒôÓà¹Ô¹§Ò¹·Õ ¼èÒ¹ÁÒ¨Ðà¡Ô´»ÃÐâª¹ì µèÍ ¤¹ ·Ó§Ò¹ à¾Õ§ ¡ÅØèÁ Ë¹Ö § à·èÒ¹Ñ ¹ áµè ËÒ¡ ͧ¤ì ¤ÇÒÁ ÃÙé áÅÐ »ÃÐʺ¡Òóì àËÅèÒ¹Ñ ¹ ¶Ù¡ ÃǺÃÇÁ áÅÐ à¼Âá¾Ãè ÊÙè Êѧ¤Áã¹Ç§¡ÇéÒ§ â´Â੾ÒÐÍÂèÒ§ÂÔ § ¡ÒèѴ·Óà» ¹ ˹ѧÊ×Í/µÓÃÒ «Ö § ·Ø¡¤¹ÊÒÁÒöÈÖ¡ÉÒ àÃÕ¹ÃÙé áÅФ鹤ÇéÒàÃ× Í§ÃÒǵèÒ§æ à¾Ô ÁàµÔÁ ÂèÍÁà·èҡѺÇèÒͧ¤ì¤ÇÒÁÃÙé´Ñ§¡ÅèÒǨСèÍãËéà¡Ô´»ÃÐ⪹ìáÅÐ ¼Å¡ÃзºµèÍÊѧ¤Áâ´ÂÃÇÁµèÍä» ÍÂèÒ§äáçµÒÁ¡ÒÃà¢Õ¹˹ѧÊ×Í/µÓÃÒäÁèãªèàÃ× Í§§èÒ à¾ÃÒмÙéà¢Õ¹µéͧ·ØèÁà··Ñ §àÇÅÒáÅФÇÒÁÁҹР¾ÂÒÂÒÁÍÂèÒ§Áҡ㹡ÒÃÃǺÃÇÁ¤ÇÒÁ¤Ô´áÅТéÍÁÙŵèÒ§æ ¨¹ÊÒÁÒöà¢Õ¹໠¹Ë¹Ñ§Ê×Í/µÓÃÒä´éÊÑ¡àÅèÁ
Ë¹Ö § ¼Á¨Ö§¢Íª× ÁªÁ㹤ÇÒÁÁØè§ÁÑ ¹¢Í§¼Ùéà¢Õ¹ ¤×Í ´Ã.» ÂÐ â¤ÇÔ¹·ì·ÇÕÇѲ¹ì «Ö §à» ¹¼Ùé·Õ ÁÕ¤ÇÒÁÃÙé ¤ÇÒÁ ÊÒÁÒö áÅФÇÒÁªÓ¹Ò·Ò§´éÒ¹Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ìà» ¹ÍÂèÒ§´Õ â´Â ´Ã.» ÂÐ â¤ÇÔ¹·ì·ÇÕÇѲ¹ì ä´é·ØèÁà·àÇÅÒáÅФÇÒÁÊÒÁÒö¨¹à¢Õ¹˹ѧÊ×ÍàÃ× Í§ ¡ÒûÃÐÁÇżÅÊÑÒ³ ÊÓËÃѺ¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅ àÅèÁ 1 áÅÐàÅèÁ 2 ¹Õ ä´éÊÓàÃ稾ÃéÍÁ¡Ñ¹·Õà´ÕÂÇ·Ñ § 2 àÅèÁ ¹Í¡¨Ò¡¹Õ ¼Á Âѧ¢Íª× Áª¹ã¹à¨µ¹ÒÃÁ³ì¢Í§ ´Ã.» ÂÐ â¤ÇÔ¹·ì·ÇÕÇѲ¹ì «Ö §¹Í¡¨Ò¡µéͧ¡ÒöèÒ·ʹͧ¤ì¤ÇÒÁÃÙéáÅÐ »ÃÐʺ¡ÒÃ³ì ¡ÒÃÃèÇÁ§Ò¹¡Ñº ÍصÊÒË¡ÃÃÁÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ã¹µèÒ§»ÃÐà·ÈáÅéÇ ¡ÒÃà¢Õ¹˹ѧÊ×Í ·Ñ § 2 àÅèÁ¹Õ ´Ã.» ÂÐ ÂѧÁØè§à¹é¹¡ÒöèÒ·ʹ´éÇÂÀÒÉÒ áÅÐÇÔ¸Õ¡ÒÃ·Õ à¢éÒ㨧èÒ à¾× ÍãËé¼ÙéÍèÒ¹·Ø¡·èÒ¹ÊÒÁÒö ·Ó¤ÇÒÁà¢éÒã¨à¹× ÍËҢͧ˹ѧÊ×Í ·Ñ § 2 àÅèÁ ä´é ÍÂèÒ§¤Ãº¶éǹáÅÐÊÁºÙóì áÅÐàª× Í ä´é ÇèÒ Í§¤ì ¤ÇÒÁÃÙé ¨Ò¡Ë¹Ñ§Ê×Í·Ñ § 2 àÅèÁ¹Õ ¨ÐªèÇ¡ÃеØé¹áÅÐÊè§àÊÃÔÁãËéÁÕ¡ÒÃÈÖ¡ÉÒ ÇԨѠáÅФ鹤ÇéÒã¹ÈÒʵÃì´éÒ¹¡Òà »ÃÐÁÇżÅÊÑÒ³ãËéá¾ÃèËÅÒÂÂÔ §¢Ö ¹µèÍä»
´Ã.¾Ñ¹¸ìÈÑ¡´Ô ÈÔÃÔÃѪµ¾§Éì ¼ÙéÍӹǡÒà ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìáÅФÍÁ¾ÔÇàµÍÃìáËè§ªÒµÔ 3 Á¡ÃÒ¤Á 2550
¤Ó¹Ó
ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì (hard disk drive) à» ¹ ¼ÅÔµÀѳ±ì ·Ò§ÍÔàÅç¡·Ã͹ԡÊì ·Õ ÁÕ ¤ÇÒÁÊÓ¤Ñ µèÍ ÀÒ¤¡ÒÃÊè§ ÍÍ¡¢Í§»ÃÐà·Èä·Â ¨Ò¡Ê¶ÔµÔ¢Í§¤³Ð¡ÃÃÁ¡ÒÃÊè§àÊÃÔÁ¡ÒÃŧ·Ø¹ (BOI) ã¹» 2547 ¾ºÇèÒ »ÃÐà·È ä·ÂÊÒÁÒöÊè§ÍÍ¡¢Í§ÍØ»¡Ã³ìÎÒÃì´´ÔÊ¡ìä´Ã¿ìáÅÐÊèǹ»ÃСͺä´éÊÙ§¶Ö§ 483,000 ÅéÒ¹ºÒ· áÅСèÍ ãËéà¡Ô´¡ÒèéÒ§§Ò¹ÁÒ¡¡ÇèÒ 100,000 ÍѵÃÒ ´Ñ§¹Ñ ¹ à¾× Íà» ¹¡ÒÃÊÃéÒ§¤ÇÒÁà¢éÁá¢ç§ãËé¡ÑºÍصÊÒË¡ÃÃÁ ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ã¹»ÃÐà·È «Ö § »ÃСͺ仴éǺÃÔÉÑ· ¼Ùé¼ÅÔµ ªÑ ¹¹Ó¢Í§âÅ¡ 4 ÃÒ ¤×Í ºÃÔÉÑ· «Õà¡· à·¤â¹âÅÂÕ (»ÃÐà·Èä·Â) ¨Ó¡Ñ´, ºÃÔÉÑ· àÇÊà·ÔÃì¹ ´Ô¨Ô·ÑÅ (»ÃÐà·Èä·Â) ¨Ó¡Ñ´, ºÃÔÉÑ· ÎÔµÒªÔ â¡ÅàºÔŠʵÍàŨ à·¤â¹âÅÂÕ ¨Ó¡Ñ´, áÅкÃÔÉÑ· ¿Ù ¨Ôµ «Ö ¡ÃзÃǧÇÔ·ÂÒÈÒʵÃì áÅÐà·¤â¹âÅÂÕ â´ÂÊӹѡ§Ò¹ ¾Ñ²¹ÒÇÔ·ÂÒÈÒʵÃìáÅÐà·¤â¹âÅÂÕáËè§ªÒµÔ (ÊÇ·ª.) ÃèÇÁ¡Ñº¼Ùéà¡Õ ÂÇ¢éͧ·Ñ §¨Ò¡ÀÒ¤àÍ¡ª¹ áÅÐʶҺѹ ¡ÒÃÈÖ¡ÉÒä´é ´Óà¹Ô¹¡ÒèѴµÑ § ¤³Ð¡ÃÃÁ¡ÒþѲ¹ÒÇÔÊÒË¡Ô¨ ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ¢Ö ¹ ÀÒÂãµé ¡ÒôÓà¹Ô¹¡Òà ¢Í§ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìáÅФÍÁ¾ÔÇàµÍÃìáËè§ªÒµÔ (๤෤) â´ÂÁÕÇѵ¶Ø»ÃÐʧ¤ìà¾× ͪèÇÂã¹ ¡ÒÃʹѺʹع áÅÐÊè§àÊÃÔÁ ¡ÒþѲ¹ÒÍصÊÒË¡ÃÃÁ´Ñ§¡ÅèÒÇã¹»ÃÐà·Èä·ÂãËé ÁÕ ¤ÇÒÁà¢éÁá¢ç§ â´Â·Ò§ ๤෤¨Ö§ ä´é ¨Ñ´àµÃÕÂÁá¼¹¡ÒÃÊè§àÊÃÔÁ ´éÒ¹µèÒ§æ äÇé ËÅÒÂâ¤Ã§¡Òà ä´éá¡è â¤Ã§¡ÒþѲ¹ÒËÅÑ¡Êٵà áÅкؤÅҡ÷ҧ´éҹ෤â¹âÅÂÕ ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì, â¤Ã§¡ÒÃÃèÇÁŧ·Ø¹ 㹡ÒÃÇԨѠáÅоѲ¹Òà·¤â¹âÅÂÕ ÎÒÃì´´ÔÊ¡ìä´Ã¿ì, áÅÐâ¤Ã§¡ÒèѴµÑ §Ê¶ÒºÑ¹ÇÔ¨ÑÂáÅоѲ¹Ò·Ò§´éÒ¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì à» ¹µé¹ ˹ѧÊ×Í ¡ÒûÃÐÁÇżÅÊÑÒ³ÊÓËÃѺ¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅ àÅèÁ 2:
¡ÒÃÍ͡Ẻǧ¨ÃÀÒ¤
ÃѺ ä´é ¶Ù¡ à¢Õ¹¢Ö ¹ ÁÒ â´ÂÁÕ ¨Ø´ÁØè§ËÁÒÂà¾× Í ãªé à» ¹ ͧ¤ì ¤ÇÒÁÃÙé ãËÁè 㹡ÒÃʹѺʹع â¤Ã§¡ÒþѲ¹Ò ´éÒ¹µèÒ§æ à¾× ÍÊè§àÊÃÔÁÍصÊÒË¡ÃÃÁÎÒÃì´´ÔÊ¡ìä´Ã¿ì¢Í§»ÃÐà·Èä·Â ˹ѧÊ×ÍàÅèÁ¹Õ àËÁÒÐÊÓËÃѺ¼Ùé·Õ ÁÕ ¾× ¹°Ò¹·Ò§´éÒ¹¡ÒûÃÐÁÇżÅÊÑÒ³´Ô¨Ô·ÑÅ (digital signal processing) áÅÐÃкºÊ× ÍÊÒôԨԷÑÅ (digital communication) ÃÇÁ·Ñ §¼Ùé·Õ ʹ㨷ҧ´éÒ¹Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì
à¹× ͧ¨Ò¡ à¹× ÍËÒÀÒÂã¹Ë¹Ñ§Ê×ÍàÅèÁ¹Õ ¨ÐµèÍà¹× ͧÁҨҡ˹ѧÊ×Í ¡ÒûÃÐÁÇżÅÊÑÒ³ÊÓËÃѺ¡Òà ¨Ñ´ à¡çº ¢éÍÁÙÅ ´Ô¨Ô·ÑÅ àÅèÁ 1:
¾× ¹°Ò¹ ªèͧÊÑÒ³ ÍèÒ¹ à¢Õ¹ à¾ÃÒÐ©Ð¹Ñ ¹ ¼Ùéà¢Õ¹ ¢Í á¹Ð¹Ó
ãËé ¼ÙéÍèÒ¹ÈÖ¡ÉÒ˹ѧÊ×Í àÅèÁ ·Õ 1 ãËé à¢éÒ㨡è͹·Õ ¨ÐÈÖ¡ÉÒ˹ѧÊ×Í àÅèÁ ¹Õ à¾× Í·Õ ¨Ðä´é à¢éÒã¨ã¹º·àÃÕ¹ ÁÒ¡ÂÔ § ¢Ö ¹ â´Âã¹Ë¹Ñ§Ê×Í àÅèÁ¹Õ ¨Ð͸ԺÒ¶֧ ËÅÑ¡¡ÒÃÍ͡Ẻǧ¨ÃÀÒ¤ÃѺ ¢Í§Ãкº¡ÒûÃÐÁÇżŠÊÑÒ³¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ã¹ÃٻẺ·Õ §èÒµèÍ ¡ÒÃà¢éÒã¨áÅÐÈÖ¡ÉÒ´éǵ¹àͧ ¹Í¡¨Ò¡¹Õ ¼Ùéà¢Õ¹ Âѧ ä´é ãªé ˹ѧÊ×Í àÅèÁ ¹Õ »ÃСͺ¡ÒÃÊ͹ÇÔªÒ ¡ÒûÃÐÁÇżÅÊÑÒ³ÊÓËÃѺ ¡ÒèѴ à¡çº ¢éÍÁÙÅ ´Ô¨Ô·ÑÅ (signal processing for digital data storage) «Ö §ã¹¢³Ð¹Õ ä´éà» ´Ê͹ã¹ËÅÑ¡ÊÙµÃÇÔÈÇ¡ÃÃÁÈÒʵà ºÑ³±Ôµ ªÑ ¹» ·Õ 4 ¢Í§ÁËÒÇÔ·ÂÒÅÑ ÃÒªÀѯ ¹¤Ã»°Á áÅÐʶҺѹ à·¤â¹âÅÂÕ ¾ÃШÍÁà¡ÅéÒ ¾Ãй¤Ã à˹×Í
¢éÒ¾à¨éÒä´éà¢Õ¹˹ѧÊ×ÍàÅèÁ¹Õ ¢Ö ¹ÁÒâ´ÂÍÒÈÑ»ÃÐʺ¡Òóì·Ñ §ËÁ´·Õ ä´éÃѺµÑ §áµèàÃÔ Á¡ÒÃÈÖ¡ÉÒÃдѺ ªÑ ¹»ÃÔÒàÍ¡·Õ Georgia Institute of Technology »ÃÐà·ÈÊËÃÑ°ÍàÁÃÔ¡Ò ¹Í¡¨Ò¡¹Õ ¢éÒ¾à¨éÒÂѧà¤Â ·Ó§Ò¹ÇԨѠ·Ò§´éÒ¹Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ³ ÈÙ¹ÂìÇԨѠ¢Í§«Õà¡· (Sea gate research center) ·Õ àÁ×ͧ Pittsburgh »ÃÐà·ÈÊËÃÑ°ÍàÁÃÔ¡Ò à» ¹ ÃÐÂÐàÇÅÒ 1 » ·ÓãËé ¢éÒ¾à¨éÒ ÁÕ ¤ÇÒÁÃÙé áÅФÇÒÁªÓ¹Ò·Ò§´éÒ¹Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì â´Â㹺··Õ 1 ¨Ð¡ÅèÒǶ֧ ¾× ¹°Ò¹·Ñ Ç仢ͧ¡Òúѹ·Ö¡ ÃкºààÁè àËÅç¡ à຺´Ô¨Ô·ÑÅ, ¡Ãкǹ¡ÒÃà¢Õ¹áÅÐÍèÒ¹¢éÍÁÙÅ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì, áÅÐà຺¨ÓÅͧªèͧÊÑÒ³¡Òúѹ·Ö¡ ÃкºààÁè àËÅç¡ «Ö § ¨Ð¶Ù¡ ¹Óä»ãªé 㹡Òà Í͡Ẻǧ¨ÃÀÒ¤ÃѺ ¢Í§Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì º··Õ 2 ¨Ð͸ԺÒ¶֧ ¤ÇÒÁÊÓ¤Ñ, ËÅÑ¡¡Ò÷ӧҹ, áÅСÒÃÍ͡Ẻ ¢Í§Ãкºä·ÁÁÔ § ÃԤѿàÇÍÃÕ (timing recovery) º· ·Õ 3 ¨Ð¡ÅèÒǶ֧ ¢Ñ ¹µÍ¹¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ (target) ààÅÐÍÕ¤ÇÍäÅà«ÍÃì (equalizer) ·Õ ãªé 㹪Ի ªèͧÊÑÒ³ÍèÒ¹ (read channel chip) º··Õ 4 ¨Ð͸ԺÒ¶֧ËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒ PRML (partial response maximum likelihood) º··Õ 5 ¨Ð¡ÅèÒǶ֧¡ÒÃÇÔà¤ÃÒÐËìà˵ءÒóì¢éͼԴ¾ÅÒ´ (er ror event) à¾× Í·Õ ¨Ð¹Ó¼ÅÅѾ¸ì ·Õ ä´é ÁÒãªé »ÃѺ»Ãا »ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ ÎÒÃì´´ÔÊ¡ì ä´Ã¿ì º··Õ 6 áÅк··Õ 7 ¨Ð͸ԺÒ¶֧ËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒ NPML (noise predictive maximum likelihood) áÅÐǧ¨ÃµÃǨËÒ PDNP (pattern dependent noise predictive) µÒÁÅӴѺ «Ö §à» ¹Ç§¨ÃµÃǨËÒ·Ñ §ÊͧẺ¹Õ ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡¡ÇèÒǧ¨ÃµÃǨËÒ PRML â´Â੾ÒÐ ÍÂèÒ§ÂÔ § àÁ× Í ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ÁÕ ¤ÇÒÁ¨Ø ÁÒ¡ áÅк··Õ 8 «Ö § à» ¹ º·ÊØ´·éÒ¢ͧ˹ѧÊ×Í àÅèÁ ¹Õ ¨ÐáÊ´§
¢Ñ ¹µÍ¹¡ÒÃÍÍ¡à຺ÃËÑÊ RLL (run length limited) «Ö § à» ¹ ÃËÑÊ ÁÍ´Ù àŪѹ ·Õ ¹ÔÂÁãªé ã¹ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ˹ѧÊ×ÍàÅèÁ¹Õ ¨ÐäÁèÊÒÁÒö·ÓãËéÊÓàÃç¨¢Ö ¹ÁÒä´éàŶéÒËÒ¡¢Ò´ºØ¤¤ÅµèÒ§æ ·Õ ¤ÍÂãËé¤ÇÒÁªèÇÂàËÅ×Í áÅÐà» ¹¡ÓÅѧã¨ãËé¢éÒ¾à¨éÒµÅÍ´ÁÒ
¢éÒ¾à¨éҢ͡ÃÒº¢Íº¾ÃФسÍÒ¨ÒÃÂì·Ø¡·èÒ¹·Õ ãËé¤ÇÒÁÃÙéáÅФÓ
»ÃÖ¡ÉÒµÅÍ´ÃÐÂÐàÇÅÒ¡ÒÃÈÖ¡ÉÒ â´Â੾ÒÐÍÂèÒ§ÂÔ § Prof. John R. Barry áÅÐ Prof. Steve W. McLaughlin ÃÇÁ·Ñ §¹Ñ¡ÇԨѨҡÈÙ¹ÂìÇԨѢͧ«Õà¡· àªè¹ Dr. Erozan M. Kurtas, Dr. M. Fatih Er den, áÅÐ Dr. Inci Ozgunes ·Õ ãËéâÍ¡ÒÊ¢éÒ¾à¨éÒä´é·Ó§Ò¹ÇԨѷҧ´éÒ¹Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì áÅÐ·Õ äÁè ÊÒÁÒö¨ÐÅ×Á ä´é ¡ç ¤×Í ·Ø¡æ ¤¹ã¹¤Ãͺ¤ÃÑÇ ¢Í§¢éÒ¾à¨éÒ ä´éá¡è ¤Ø³ à¡ÕÂÃµÔ â¤ÇÔ¹·ì·ÇÕÇѲ¹ì, ¤Ø³¾ÃÃ³Õ â¤ÇÔ¹·ì·ÇÕÇѲ¹ì, ¤Ø³ÃѪ¹ÔÈ â蹡Ԩ, ¤Ø³Í¹Ø·Ñȹì â蹡Ԩ, ¤Ø³ ©ÑµÃªÑ â¤ÇÔ¹·ì·ÇÕÇѲ¹ì, ¤Ø³¡ÔµµÔÈÑ¡´Ô â¤ÇÔ¹·ì·ÇÕÇѲ¹ì, áÅФسÈÔÃÊØ´Ò âÊÁ¹ÑÊ ¹Í¡¨Ò¡¹Õ ¢éÒ¾à¨éÒ¢Í ¢Íº¤Ø³ ÁËÒÇÔ·ÂÒÅÑ ÃÒªÀѯ ¹¤Ã»°Á·Õ ãËé¡ÒÃʹѺʹع áÅÐãËé ¤ÇÒÁÊдǡ¢éÒ¾à¨éÒ µÅÍ´ÃÐÂÐàÇÅÒ ã¹¡ÒÃà¢Õ¹˹ѧÊ×ÍàÅèÁ¹Õ ·éÒÂÊØ´¹Õ ¢éÒ¾à¨éÒä´é¾ÂÒÂÒÁÍÂèÒ§ÂÔ §ã¹¡ÒÃ·Õ ¨Ð·ÓãËé˹ѧÊ×ÍàÅèÁ¹Õ §èÒµèÍ¡ÒÃàÃÕ¹ÃÙé à¾× ÍãËé¼ÙéÍèÒ¹ ÊÒÁÒö·Ó¤ÇÒÁà¢éÒã¨ä´é ´éǵ¹àͧÍÂèÒ§ÃÇ´àÃçÇ áÅÐÁÕ »ÃÐÊÔ·¸Ô¼Å Ëҡ˹ѧÊ×Í àÅèÁ ¹Õ ÁÕ ¢éͺ¡¾Ãèͧ »ÃСÒÃã´ ¢éÒ¾à¨éÒ ÁÕ ¤ÇÒÁÂÔ¹´Õ áÅШѡ ¢Íº¾ÃФس ÂÔ § ËÒ¡·èÒ¹¼Ùéãªé ˹ѧÊ×Í àÅèÁ ¹Õ ¨ÐÊè§ ¢éͤԴàËç¹ áÅÐ ¤Óá¹Ð¹Ó·Õ à» ¹»ÃÐ⪹ìÊÓËÃѺ¡ÒûÃѺ»Ãا˹ѧÊ×ÍàÅèÁ¹Õ ÁÒ·Õ ÍÕàÁÅì piya@npru.ac.th à¾× Í·Õ ¢éÒ¾à¨éÒ ¨Ðä´é ´Óà¹Ô¹¡ÒûÃѺ»Ãا áÅÐá¡éä¢ã¹¡ÒþÔÁ¾ì ¤ÃÑ § µèÍä» ÊÓËÃѺ ¢éÍÁÙÅ ¢èÒÇÊÒõèÒ§æ à¡Õ ÂǡѺ ˹ѧÊ×Í àÅèÁ¹Õ áÅÐÍ× ¹æ ÊÒÁÒöµÔ´µÒÁä´é·Õ àÇçºä«µì http://home.npru.ac.th/∼t3058
´Ã.» ÂÐ â¤ÇÔ¹·ì·ÇÕÇѲ¹ì â»Ãá¡ÃÁÇÔÈÇ¡ÃÃÁâ·Ã¤Á¹Ò¤Á ÁËÒÇÔ·ÂÒÅÑÂÃÒªÀѯ¹¤Ã»°Á Á¡ÃÒ¤Á 2550
ÊÒúÑ
1
2
º·¹Ó
1
1.1
¾× ¹°Ò¹¢Í§¡Òúѹ·Ö¡ÃкºààÁèàËÅç¡à຺´Ô¨Ô·ÑÅ . . . . . . . . . . . . . . . . . . .
1
1.2
à຺¨ÓÅͧ¢Í§Ãкº¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì . . . . . . . . . . . .
3
1.3
¡Ãкǹ¡ÒÃà¢Õ¹
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
1.4
¡Ãкǹ¡ÒÃÍèÒ¹ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
1.5
à຺¨ÓÅͧªèͧÊÑÒ³¡Òúѹ·Ö¡ÃкºààÁèàËÅç¡ . . . . . . . . . . . . . . . . . .
9
1.5.1
à຺¨ÓÅͧªèͧÊÑÒ³àÊÁ×͹¨ÃÔ§
. . . . . . . . . . . . . . . . . . . .
11
1.5.2
à຺¨ÓÅͧªèͧÊÑÒ³ÍØ´Á¤µÔ
. . . . . . . . . . . . . . . . . . . . . .
12
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15
1.6
ÊÃØ»·éÒº·
1.7
à຺½ ¡ËÑ´·éÒº·
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15
ä·ÁÁÔ §ÃԤѿàÇÍÃÕ
17
2.1
º·¹Ó
17
2.2
ä·ÁÁÔ §ÃԤѿàÇÍÃÕà຺·Õ ãªé¡Ñ¹·Ñ Çä»
2.3
¡ÒÃÍÍ¡à຺¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL
2.4
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
19
. . . . . . . . . . . . . . . . . . . . .
23
2.3.1
¡ÒÃÇÔà¤ÃÒÐËìàªÔ§àÊ鹢ͧǧ¨Ã PLL Íѹ´Ñº·Õ Ë¹Ö § . . . . . . . . . . . . . .
23
2.3.2
¡ÒÃÇÔà¤ÃÒÐËìàªÔ§àÊ鹢ͧǧ¨Ã PLL Íѹ´Ñº·Õ Êͧ
. . . . . . . . . . . . . .
28
2.3.3
¡ÒÃËÒàÊé¹â¤é§ÃÙ»µÑÇàÍÊ
. . . . . . . . . . . . . . . . . . . . . . . . . .
30
»ÃÐÊÔ·¸ÔÀÒ¾¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕà຺·Õ ãªé¡Ñ¹·Ñ Çä» . . . . . . . . . . . . . . . . .
34
ix
3
4
2.5
ä·ÁÁÔ §ÃԤѿàÇÍÃÕà຺´Ô¨Ô·ÑÅ . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
37
2.6
àà¹Çâ¹éÁ¢Í§Ãкºä·ÁÁÔ §ÃԤѿàÇÍÃÕã¹Í¹Ò¤µ . . . . . . . . . . . . . . . . . . . .
38
2.7
ÊÃØ»·éÒº·
41
2.8
à຺½ ¡ËÑ´·éÒº·
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
41
¡ÒÃÍÍ¡à຺·ÒÃìà¡çµààÅÐÍÕ¤ÇÍäÅà«ÍÃì
43
3.1
º·¹Ó
43
3.2
¡ÒÃÍÍ¡à຺·ÒÃìà¡çµ´éÇÂÇÔ¸Õ¡Òà MMSE
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.1
à§× ͹䢺ѧ¤Ñºà຺âÁ¹Ô¡ (h0
3.2.2
à§× ͹䢺ѧ¤Ñºà຺
3.2.3
à§× ͹䢺ѧ¤Ñºà຺¾Åѧ§Ò¹Ë¹Ö §Ë¹èÇ (H
. . . . . . . . . . . .
53
3.2.4
à§× ͹䢺ѧ¤Ñºà຺·ÒÃìà¡çµà©¾ÒÐ . . . . . . . . . . . . . . . . . . . . . .
53
3.3
¼Å¡Ò÷´Åͧ
3.4
ÊÃØ»·éÒº·
3.5
à຺½ ¡ËÑ´·éÒº·
h1 = 1
= 1)
47
. . . . . . . . . . . . . . . . . . . . .
49
. . . . . . . . . . . . . . . . . . . . . . . . .
52
TH
= 1)
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
54
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
62
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
64
ǧ¨ÃµÃǨËÒ PRML
65
4.1
º·¹Ó
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
65
4.2
ÍÕ¤ÇÍäÅà«ÍÃì . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
66
4.2.1
ÍÕ¤ÇÍäÅà«ÍÃìà຺¼ÅµÍºÊ¹Í§àµçÁ
. . . . . . . . . . . . . . . . . . . .
67
4.2.2
ÍÕ¤ÇÍäÅà«ÍÃìà຺¼ÅµÍºÊ¹Í§ºÒ§Êèǹ . . . . . . . . . . . . . . . . . .
68
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
69
4.3.1
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
71
4.3.2
à༹ÀÒ¾à·ÃÅÅÔÊ
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
72
4.3.3
ÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
75
4.3.4
¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ . . . . . . . . . . . . . . . . . . .
80
µÑÇÍÂèÒ§¡ÒÃãªé§Ò¹Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ . . . . . . . . . . . . . . . . . . . . . . .
81
4.3
4.4
4.4.1
5
6
7
ÊÃػǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
. . . . . . . . . . . . . . . . . . . . . . . . .
87
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
4.5
ÊÃØ»·éÒº·
4.6
à຺½ ¡ËÑ´·éÒº·
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
89
¡ÒÃÇÔà¤ÃÒÐËìà˵ءÒóì¢éͼԴ¾ÅÒ´
91
5.1
º·¹Ó
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
91
5.2
¤ÇÒÁËÁÒ¢ͧà˵ءÒóì¢éͼԴ¾ÅÒ´ . . . . . . . . . . . . . . . . . . . . . . . .
93
5.3
ÃÐÂзҧÂؤÅÔ´ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
97
5.4
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
99
5.5
¼Å¡Ò÷´Åͧ
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
5.5.1
¡ÒÃÇÔà¤ÃÒÐËìÃÐÂзҧ·Õ ¹éÍÂÊØ´ . . . . . . . . . . . . . . . . . . . . . . . 104
5.5.2
¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§
5.6
ÊÃØ»·éÒº·
5.7
à຺½ ¡ËÑ´·éÒº·
SNReff
ààÅÐ BER
. . . . . . . . . . . . . . . . . 107
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
ǧ¨ÃµÃǨËÒ NPML
115
6.1
º·¹Ó
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
6.2
¡Ãкǹ¡ÒÃ㹡Ò÷ӹÒÂÊÑҳú¡Ç¹ . . . . . . . . . . . . . . . . . . . . . 117
6.3
¡ÒÃËÒ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂ
6.4
ËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒ NPML . . . . . . . . . . . . . . . . . . . . . . 120
6.5
¼Å¡Ò÷´Åͧ
6.6
ÊÃØ»·éÒº·
6.7
à຺½ ¡ËÑ´·éÒº·
. . . . . . . . . . . . . . . . . . . . . 118
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
ǧ¨ÃµÃǨËÒ PDNP
137
7.1
º·¹Ó
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
7.2
¡ÒÃ¢Ö ¹ÍÂÙè¡Ñºà຺¢éÍÁÙŢͧÊÑҳú¡Ç¹
7.3
ÍÑÅ¡ÍÃÔ·ÖÁ PDNP
. . . . . . . . . . . . . . . . . . . . 138
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
8
7.4
ÍÑÅ¡ÍÃÔ·ÖÁ PS PDNP
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
7.5
¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒ PDNP
7.6
¼Å¡Ò÷´Åͧ
7.7
ÊÃØ»·éÒº·
7.8
à຺½ ¡ËÑ´·éÒº·
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
¡ÒÃÍÍ¡à຺ÃËÑÊ RLL
151
. . . . . . . . . . . . . . . . . . . . . . . 146
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
8.1
º·¹Ó
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
8.2
¨Ó¹Ç¹ÅӴѺ¢éÍÁÙÅ·Ñ §ËÁ´·Õ ÊÍ´¤Åéͧ¡Ñºà§× ͹䢺ѧ¤Ñº
8.3
¤ÇÒÁ¨Ø¢Í§ÃËÑÊ RLL à຺
(d, k)
(d, k)
. . . . . . . . . . . 153
. . . . . . . . . . . . . . . . . . . . . . . . . 154
(d, k)
8.3.1
ÍѵÃÒ¢èÒÇÊÒÃàªÔ§àÊ鹡ӡѺ¢Í§ÃËÑÊ RLL à຺
. . . . . . . . . . . 155
8.3.2
ÍѵÃÒ¤ÇÒÁ˹Òàà¹è¹ . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
8.4
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL . . . . . . . . . . . . . . . . . . . . . . . . . . 156
8.5
àÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
8.5.1
¡ÒÃËÒÍѵÃÒ¢èÒÇÊÒÃàªÔ§àÊ鹡ӡѺ . . . . . . . . . . . . . . . . . . . . . . 160
8.5.2
ÅӴѺ¢éÍÁÙÅ·Õ ÊÍ´¤Åéͧ¡Ñºà¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL à຺
(d, k)
. . 160
8.6
¢Ñ ¹µÍ¹¡ÒÃÍÍ¡à຺ÃËÑÊ RLL . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
8.7
µÑÇÍÂèÒ§ÃËÑÊ RLL à຺µèÒ§æ . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
8.8
ÃËÑÊ
8.9
ÊÃØ»·éÒº·
(0, G/I)
ÊÓËÃѺªèͧÊÑÒ³ PRML
. . . . . . . . . . . . . . . . . . . . 166
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
8.10 à຺½ ¡ËÑ´·éÒº·
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Q
¡
µÒÃÒ§¿ §¡ìªÑ¹
¢
Êٵä³ÔµÈÒʵÃì·Õ ÊÓ¤Ñ
171
175
¢.1
µÃÕ⡳ÁÔµÔ (Trigonometric)
. . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
¢.2
»ÃԾѹ¸ìäÁè¨Ó¡Ñ´à¢µ (Inde nite Integral) . . . . . . . . . . . . . . . . . . . . . . 176
¤
¡ÒÃËÒ͹ؾѹ¸ì¢Í§àÇ¡àµÍÃìààÅÐàÁ·ÃÔ¡«ì
177
§
¤ÓÈѾ·ìà·¤¹Ô¤
179
ºÃóҹءÃÁ
191
´Ãê¹Õ
199
ÊÒúÑÃÙ»
1.1
ËÅÑ¡¡ÒÃ¾× ¹°Ò¹¢Í§¡Òúѹ·Ö¡ÃкºáÁèàËÅç¡ . . . . . . . . . . . . . . . . . . . .
2
1.2
Ẻ¨ÓÅͧ·Ñ Ç仢ͧÃкº¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì [6] . . . . . . .
4
1.3
ÊÑÒ³¾ÑÅÊì à»ÅÕ Â¹Ê¶Ò¹Ð ÊÓËÃѺ ¡Òúѹ·Ö¡ (a) Ẻá¹Ç¹Í¹ áÅÐ (b) Ẻ á¹ÇµÑ § . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7
1.4
¼ÅµÍºÊ¹Í§ä´ºÔµ ÊÓËÃѺ¡Òúѹ·Ö¡ (a) Ẻá¹Ç¹Í¹ áÅÐ (b) Ẻá¹ÇµÑ § . . .
9
1.5
¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§ÊÑÒ³¾ÑÅÊìä´ºÔµ ÊÓËÃѺ ¡Òúѹ·Ö¡ (a) Ẻá¹Ç ¹Í¹ áÅÐ (b) Ẻá¹ÇµÑ § . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10
1.6
Ẻ¨ÓÅͧªèͧÊÑÒ³àÊÁ×͹¨ÃÔ§ . . . . . . . . . . . . . . . . . . . . . . . . .
11
1.7
Ẻ¨ÓÅͧªèͧÊÑÒ³ÍØ´Á¤µÔ
13
1.8
¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§·ÒÃìà¡çµ ẺµèÒ§æ ÊÓËÃѺ ªèͧÊÑÒ³¡Òúѹ·Ö¡ (a)
. . . . . . . . . . . . . . . . . . . . . . . . . .
Ẻá¹Ç¹Í¹ áÅÐ (b) Ẻá¹ÇµÑ §
. . . . . . . . . . . . . . . . . . . . . . . .
14
2.1
ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ¹ÔùÑ . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
19
2.2
Ẻ¨ÓÅͧªèͧÊÑÒ³ÍØ´Á¤µÔ ¾ÃéÍÁ¡Ñºä·ÁÁÔ §ÃԤѿàÇÍÃÕẺÍØ»¹Ñ . . . . . . . .
19
2.3
¤èÒÁÒ¡ÊØ´¢Í§
2.4
(a) ¤èÒ
αC
áµèÅÐ
d,
2.5
α
·Õ Âѧ¤§·ÓãËéÃкºÁÕ¤ÇÒÁàʶÕÂÃÀÒ¾ ÊÓËÃѺáµèÅФèÒ
·Õ ·ÓãËéÃкºÁÕ¤ÇÒÁàʶÕÂÃÀÒ¾áÅÐÁÕÍѵÃÒ¡ÒÃÅÙèà¢éÒÀÒÂã¹
(a) ¼ÅµÍºÊ¹Í§¢Ñ ¹ºÑ¹ä´¢Í§Ãкº áÅÐ (b) ¼ÅµÍºÊ¹Í§¢éͼԴ¾ÅÒ´ÊÓËÃѺ
d=
14T
áÅФèÒ
α
µèÒ§æ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
25
ºÔµ ÊÓËÃѺ
30T
ÊÓËÃѺ¤èÒ
d
C
. . . . . .
¨Ò¡ 0 ¶Ö§
áÅÐ (b) ¼ÅµÍºÊ¹Í§¢Í§ÃкºàÁ× Íãªé
α100
d
26
27
E(z)
ºÔµ àÁ× Íãªé
d = 14
áÅÐ
αC
¢¹Ò´ÁÒ¡ÊØ´¢Í§
2.7
Ẻ¨ÓÅͧÊÓËÃѺ¡ÒÃËÒàÊé¹â¤é§ÃÙ»µÑÇàÍÊ
2.8
àÊé¹â¤é§ ÃÙ» µÑÇ àÍʢͧǧ¨Ã M&M TED ÊÓËÃѺ ªèͧÊÑÒ³ PR4 ·Õ ãªé ä·ÁÁÔ §
2.9
ËÅѧ¨Ò¡·Õ ¢éÍÁÙżèÒ¹ä»
C
2.6
. .
29
. . . . . . . . . . . . . . . . . . . . .
31
ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä» . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
32
µÑÇÍÂèÒ§ÅѡɳТͧä«à¤ÔÅÊÅÔ» . . . . . . . . . . . . . . . . . . . . . . . . . . .
33
2.10 (a) ¢éͼԴ¾ÅÒ´·Ò§àÇÅÒẺ RMS
σ² /T
ªèͧÊÑÒ³ÍØ´Á¤µÔẺ PR4 ·Õ ÁÕ¤èÒ ¤ÇÒÁ¶Õ )
áÅÐ (b) »ÃÐÊÔ·¸ÔÀÒ¾¢Í§ BER ÊÓËÃѺ
σw /T
µèÒ§æ ¡Ñ¹ (àÁ× ÍÃкºäÁèÁÕÍͿ૵·Ò§
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
35
2.11 »ÃÐÊÔ·¸ÔÀÒ¾ BER ¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Ç仢ͧªèͧÊÑÒ³ÍØ´Á¤µÔ Ẻ PR4 ÊÓËÃѺ
σw /T = 0.5%
áÅÐÍͿ૵·Ò§¤ÇÒÁ¶Õ 0.2%
. . . . . . . . .
36
2.12 â¤Ã§ÊÃéÒ§¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ»ÃÐÁÒ³¤èÒ㹪èǧ . . . . . . . . . . . . . . . .
38
2.13 »ÃÐÊÔ·¸ÔÀÒ¾¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺµèÒ§æ ¢Í§ªèͧÊÑÒ³ÍØ´Á¤µÔẺ PR4 . . .
39
2.14 ÍѵÃÒ¡ÒÃÅÙèà¢éҢͧä·ÁÁÔ §ÃԤѿàÇÍÃÕẺµèÒ§æ àÁ× Íãªé
3.1
α50
·Õ
Eb /N0 = 10
dB
. . .
40
¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§·ÒÃìà¡çµ ẺµèÒ§æ ÊÓËÃѺ ÃкººÑ¹·Ö¡ (a) Ẻá¹Ç ¹Í¹ áÅÐ (b) Ẻá¹ÇµÑ § . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
46
3.2
Ẻ¨ÓÅͧÊÓËÃѺ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ´éÇÂÇÔ¸Õ¡Òà MMSE . . . . . . . . . . . . .
48
3.3
»ÃÐÊÔ·¸ÔÀÒ¾ã¹ÃÙ» BER ¢Í§Ãкº·Õ ãªé·ÒÃìà¡çµµèÒ§æ ÊÓËÃѺ ND = 2
. . . . . . .
59
3.4
»ÃÐÊÔ·¸ÔÀÒ¾ã¹ÃÙ» BER ¢Í§Ãкº·Õ ãªé·ÒÃìà¡çµµèÒ§æ ·Õ ND = 2.5 . . . . . . . . .
60
3.5
¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ ¢Í§·ÒÃìà¡çµáººµèÒ§æ à·Õº¡ÑºªèͧÊÑÒ³·Õ ND = 2.5
61
3.6
ÍѵÊËÊÑÁ¾Ñ¹¸ì¢Í§ÅӴѺ¢éÍÁÙÅ
3.7
(a) ¡ÃÒ¿ÃÐËÇèÒ§ SNR ·Õ µéͧ¡Òà áÅÐ ND àÁ× Í SNR ·Õ µéͧ¡Òà áÅÐ
σj
{wk }
.
ÊÓËÃѺÃкº·Õ ·ÒÃìà¡çµáººµèÒ§æ ·Õ ND = 2.5
σj = 0%
62
áÅÐ (b) ¡ÃÒ¿ÃÐËÇèÒ§
·Õ ND = 2.5 . . . . . . . . . . . . . . . . . . . . . . .
4.1
ËÅÑ¡¡ÒÃ¾× ¹°Ò¹¢Í§à·¤¹Ô¤ PRML
. . . . . . . . . . . . . . . . . . . . . . . .
4.2
Ẻ¨ÓÅͧªèͧÊÑÒ³·Õ äÁèµèÍà¹× ͧ·Ò§àÇÅÒẺÊÁÁÙÅ
4.3
µÑÇÍÂèҧẺ¨ÓÅͧªèͧÊÑÒ³·Õ äÁèµèÍà¹× ͧ·Ò§àÇÅÒẺÊÁÁÙÅÅѡɳеèÒ§æ
. . . . . . . . . . . . . . . . .
63
66 66 70
H(D) = 1 + D
4.4
á¼¹ÀÒ¾ªèͧÊÑҳẺ PR1,
4.5
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ªèͧÊÑÒ³ PR1,
4.6
á¼¹ÀÒ¾à·ÃÅÅÔʢͧªèͧÊÑÒ³ PR1,
4.7
á¼¹ÀÒ¾à·ÃÅÅÔʢͧªèͧÊÑÒ³ EPR4,
4.8
(a) á¼¹ÀÒ¾ºÅçÍ¡, (b) à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´, áÅÐ (c) á¼¹ÀÒ¾à·ÃÅÅÔÊ ¢Í§ªèͧ ÊÑÒ³
4.9
H(D) = 1 − D2
Ẻ¨ÓÅͧªèͧÊÑÒ³
. . . . . . . . . . . . . . . . .
H(D) = 1 + D
H(D) = 1 + D
. . . . . . . . . . . .
72
. . . . . . . . . . . . .
73
H(D) = 1 + D − D2 − D3
. . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
H(D)
¾ÃéÍÁǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
71
74
75
. . . . . . . . . . . .
76
4.10 ¤Ó͸ԺÒÂá¼¹ÀÒ¾à·ÃÅÅÔÊ . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
76
4.11 á¼¹ÀÒ¾à·ÃÅÅÔʢͧªèͧÊÑÒ³ PR4,
H(D) = 1 − D2
. . . . . . . . . . . .
77
4.12 ¢Ñ ¹µÍ¹¡Ò÷ӧҹ¢Í§ÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ . . . . . . . . . . . . . . . . . . . . . .
79
4.13 á¼¹ÀÒ¾à·ÃÅÅÔʢͧªèͧÊÑÒ³
H(D) = 1 + 0.5D
. . . . . . . . . . . . . .
81
4.14 á¼¹ÀҾ͸ԺÒ¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔã¹áµèÅЪèǧàÇÅÒ . . . . . . . .
82
4.15 (a) á¼¹ÀÒ¾ºÅçÍ¡, (b) à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´, áÅÐ (c) á¼¹ÀÒ¾à·ÃÅÅÔÊ ¢Í§ªèͧ ÊÑÒ³
H(D) = 1 − D
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
86
4.16 á¼¹ÀÒ¾ÊÃØ»¢Ñ ¹µÍ¹¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ . . . . . . . . . . . . . .
87
5.1
µÑÇÍÂèÒ§¡Ò÷ӧҹÀÒÂã¹à·ÃÅÅÔʢͧǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ . . . . . . . . . . . . .
92
5.2
Ẻ¨ÓÅͧªèͧÊÑÒ³ GPR ẺÊÁÁÙÅ
. . . . . . . . . . . . . . . . . . . . .
93
5.3
µÑÇÍÂèÒ§¡ÒäӹdzËÒÅӴѺ¢éͼԴ¾ÅÒ´
. . . . . . . . . . . . . . . . . . . . . .
95
5.4
ÀÒ¾ÊͧÁÔµÔáÊ´§¡Ãкǹ¡ÒöʹÃËÑÊ¢éÍÁÙÅâ´Âãªéǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
5.5
»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкºã¹ÃÙ» (a) BER áÅÐ (b)
5.6
(a) ¡ÃÒ¿ BER áÅÐ
SNReff ,
SNReff
. . . . . .
98
¢Í§·ÒÃìà¡çµáºº GPR5 . . 109
(b) ¡ÃÒ¿ BER áÅÐ SNR ¢Í§Ãкº·Õ ãªé ·ÒÃìà¡çµ
ẺµèÒ§æ ³ ÃдѺ¤ÇÒÁÃعáç¢Í§ÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡·Õ à¢Õ¹ÇèÒ jitter µèÒ§æ ·Õ ND = 2.5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
6.1
Ẻ¨ÓÅͧªèͧÊÑÒ³·Õ äÁèµèÍà¹× ͧ·Ò§àÇÅÒẺÊÁÁÙÅ
6.2
Ẻ¨ÓÅͧªèͧÊÑÒ³¾ÃéÍÁǧ¨ÃµÃǨËÒ NPML
. . . . . . . . . . . . . . 116
. . . . . . . . . . . . . . . . 117
6.3
Ẻ¨ÓÅͧ¡Ãкǹ¡ÒÃ㹡Ò÷ÓãËéÊÑҳú¡Ç¹à» ¹ÊÕ¢ÒÇ
6.4
Ẻ¨ÓÅͧªèͧÊÑÒ³ PR4
6.5
â¤Ã§ÊÃéÒ§¢Í§Ç§¨ÃµÃǨËÒ NPML ·Õ ãªé ¡Ñº §Ò¹»ÃÐÂØ¡µì ·Õ µéͧ¡ÒäÇÒÁàÃçÇ ã¹¡Òà »ÃÐÁÇżÅÊÙ§
. . . . . . . . . . . 118
. . . . . . . . . . . . . . . . . . . . . . . . . . . 121
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
6.6
Ẻ¨ÓÅͧªèͧÊÑҳẺÊÁÁÙÅ ¾ÃéÍÁ·Ñ §Ç§¨ÃµÃǨËÒ NPML áÅÐ PRML
6.7
á¼¹ÀÒ¾à·ÃÅÅÔʢͧ (a) ·ÒÃìà¡çµ
Heff (D)
=
H(D)
1 − 0.247D − 0.753D2
=
1−D
. . 126
áÅÐ (b) ·ÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Å
·Õ ãªé㹡ÒöʹÃËÑÊ¢éÍÁÙŢͧÃкº PRML
áÅÐ NPML µÒÁÅӴѺ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 6.8
á¼¹ÀÒ¾ÊÃØ»¢Ñ ¹µÍ¹¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔÊÓËÃѺÃкº PRML
. . . 128
6.9
á¼¹ÀÒ¾ÊÃØ»¢Ñ ¹µÍ¹¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔÊÓËÃѺÃкº NPML . . . 129
6.10 Ẻ¨ÓÅͧªèͧÊÑÒ³¢Í§Ãкº¡Òúѹ·Ö¡áÁèàËÅç¡ . . . . . . . . . . . . . . . . 130 6.11 »ÃÐÊÔ·¸ÔÀÒ¾¢Í§ÃкºµèÒ§æ ·Õ ND = 2
. . . . . . . . . . . . . . . . . . . . . . 131
6.12 »ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂ·Õ ãªé¨Ó¹Ç¹á·ç»µèÒ§¡Ñ¹ ·Õ SNR = 17 dB . . . . 132 6.13 »ÃÐÊÔ·¸ÔÀÒ¾¢Í§ÃкºµèÒ§æ ·Õ ND = 2.5
. . . . . . . . . . . . . . . . . . . . . 133
6.14 ¼Å¡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§ÃкºµèÒ§æ ·Õ ND = 2.5 . . . . . . . . . . . . 134
7.1
Ẻ¨ÓÅͧªèͧÊÑÒ³
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
7.2
¡ÓÅѧÊÑҳú¡Ç¹·Õ ¢Ö ¹¡Ñºáºº¢éÍÁÙÅ ³ ´éÒ¹¢ÒÍÍ¡¢Í§ÍÕ¤ÇÍäÅà«ÍÃì·Õ ¶Ù¡Í͡Ẻ ÊÓËÃѺ·ÒÃìà¡çµ EEPR2 [1 4 6 4 1] ¢Í§Ãкº¡Òúѹ·Ö¡áººá¹ÇµÑ § ·Õ ND = 2.5, SNR = 30 dB, áÅÐ
7.3
σj /T = 10%
. . . . . . . . . . . . . . . . . . . . . . . . 140
(a) á¼¹ÀÒ¾à·ÃÅÅÔÊÊÓËÃѺ·ÒÃìà¡çµáºº PR4 áÅÐ (b) ¢Ñ ¹µÍ¹¡ÒöʹÃËÑÊ¢éÍÁÙÅ ´éÇÂá¼¹ÀÒ¾à·ÃÅÅÔÊ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
7.4
»ÃÐÊÔ·¸ÔÀÒ¾ã¹ÃÙ» ¢Í§ BER ¢Í§Ç§¨ÃµÃǨËÒµèÒ§æ ÊÓËÃѺ Ãкº¡Òúѹ·Ö¡ (a) Ẻá¹Ç¹Í¹ áÅÐ (b) Ẻá¹ÇµÑ § ·Õ ND = 2.5 áÅÐ
σj /T = 10%
. . . . . . . 149
8.1
Ẻ¨ÓÅͧ¡ÒÃà¢éÒÃËÑÊ RLL . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
8.2
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ
(d, k)
. . . . . . . . . . . . . . . . . . . 157
8.3
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ
(1, 3)
. . . . . . . . . . . . . . . . . . . 158
8.4
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ
(0, 3)
. . . . . . . . . . . . . . . . . . . 160
8.5
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ
(0, 2)
. . . . . . . . . . . . . . . . . . . 161
8.6
µÑÇÍÂèÒ§ÃËÑÊ RLL ẺµèÒ§æ ·Õ ãªéã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì . . . . . . . . . . . . . . . . 165
ÊÒúѵÒÃÒ§
3.1
µÑÇÍÂèÒ§·ÒÃìà¡çµáºº PR ·Õ à» ¹·Õ ÂÍÁÃѺ¡Ñ¹ã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì
4.1
¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§
4.2
µÑÇÍÂèÒ§áÊ´§¨Ó¹Ç¹Ê¶Ò¹Ð·Õ µéͧãªéã¹á¼¹ÀÒ¾à·ÃÅÅÔʢͧ·ÒÃìà¡çµáººµèÒ§æ
5.1
ÃÐÂзҧÂؤÅÔ´¡ÓÅѧÊͧ
ak
áÅÐ
5.2
. . .
71 88
εa (D) ¢Í§Ãкº·Õ ãªé·ÒÃìà¡çµáºº GPR3 ·Õ ND = 2.5
εa (D)
¢Í§Ãкº·Õ ãªé ·ÒÃìà¡çµ Ẻ PR2 ·Õ ND=2.5
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Õ ÃкºÁÕ BER =
. . . . . . .
d2 {εa (D)}, ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧÊͧ d2eff {εa (D)},
áÅÐÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ
5.3
H(D) = 1 + D
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
ÃÐÂзҧÂؤÅÔ´¡ÓÅѧÊͧ
áÅÐ SNR=22 dB
¢Í§ªèͧÊÑÒ³
45
d2 {εa (D)}, ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧÊͧ d2eff {εa (D)},
áÅÐÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ áÅÐ SNR = 22 dB
rk
. . . . . . . . . . .
10−4
εa (D)
¢Í§Ãкº·Õ ãªé·ÒÃìà¡çµáÅÐ
σj /T
ẺµèÒ§æ ³ ¨Ø´
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
7.1
¨Ó¹Ç¹¢Í§µÑÇ´Óà¹Ô¹¡ÒÃ·Õ ãªéµèÍ¢éÍÁÙÅ 1 ºÔµ ¢Í§Ç§¨ÃµÃǨËÒ PDNP áÅÐ PS PDNP147
8.1
µÑÇÍÂèÒ§¨Ó¹Ç¹ÅӴѺ¢éÍÁÙÅ·Ñ §ËÁ´ ºÑ§¤Ñº
(d, ∞)
Nd (L)
·Õ ÁÕ¤ÇÒÁÂÒÇ
L
·Õ ÊÍ´¤Åéͧ¡Ñºà§× ͹ä¢
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
(d, k)
8.2
ÍѵÃÒ¢èÒÇÊÒÃàªÔ§àÊ鹡ӡѺ¢Í§ÃËÑÊ RLL Ẻ
8.3
¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§¤ÇÒÁ¨Ø
8.4
µÒÃÒ§¤é¹ËÒÊÓËÃѺ¡ÒÃà¢éÒáÅжʹÃËÑÊ RLL Ẻ
C(d, k)
µèÒ§æ
. . . . . . . . . . . . 156
áÅÐÍѵÃÒ¤ÇÒÁ˹Òá¹è¹ DR . . . . . . . . . 157
xxi
(0, 2)
. . . . . . . . . . . . . 163
º··Õ 1
º·¹Ó
㹺·¹Õ ¨Ð͸ԺÒ¶֧ÀÒ¾ÃÇÁ¢Í§Ãкº¡Òúѹ·Ö¡¢éÍÁÙŢͧÍØ»¡Ã³ìÎÒÃì´´ÔÊ¡ìä´Ã¿ì (hard disk drive) à¾× Í à» ¹ ¡ÒÃàµÃÕÂÁ¤ÇÒÁ¾ÃéÍÁãËé ¼ÙéÍèÒ¹à¢éÒ㨶֧ ¾× ¹°Ò¹µèÒ§æ ·Õ à¡Õ ÂÇ¢éͧ¡Ñº Ãкº¡Òúѹ·Ö¡ ¢éÍÁÙŠẺ´Ô¨Ô·ÑÅ ÃÇÁ·Ñ § ËÅÑ¡¡Ò÷ӧҹ¢Í§¡Ãкǹ¡ÒÃà¢Õ¹áÅСÒÃÍèÒ¹¢éÍÁÙÅ ã¹ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ¡è͹ ·Õ ͸ԺÒÂà¡Õ ÂǡѺ ¡ÒÃÇÔà¤ÃÒÐËì Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³áÅСÒÃÍ͡Ẻǧ¨ÃÀÒ¤ÃѺ ¢Í§ÎÒÃì´ ´ÔÊ¡ìä´Ã¿ì㹺·µèÍä»
1.1
¾× ¹°Ò¹¢Í§¡Òúѹ·Ö¡ÃкºààÁèàËÅç¡à຺´Ô¨Ô·ÑÅ
¡Òúѹ·Ö¡ÃкºáÁèàËÅç¡ (magnetic recording) ¤×Í ¡ÒèѴà¡çº¢éÍÁÙźԵãËéÍÂÙèã¹ÃÙ»¢Í§¡ÒÃà»ÅÕ Â¹ á»Å§ÃдѺÊÀÒ¾¤ÇÒÁà» ¹áÁèàËÅç¡ (magnetization) ã¹Ê× ÍºÑ¹·Ö¡ «Ö §ÊÒÁÒöáºè§ÍÍ¡à» ¹ 2 Ẻ [1] ¤×Í áººá͹ÐÅçÍ¡ (analog) áÅÐẺ´Ô¨Ô·ÑÅ (digital) ã¹Ë¹Ñ§Ê×Í àÅèÁ ¹Õ ¨Ð¾Ô¨ÒóÒ੾ÒСÒà ºÑ¹·Ö¡ ÃкºáÁèàËÅç¡ áºº´Ô¨Ô·ÑÅ ·Õ ãªé ÊÓËÃѺ ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì à·èÒ¹Ñ ¹ â´Â·Õ ¡Òúѹ·Ö¡ ÃкºáÁèàËÅç¡ áºº´Ô¨Ô·ÑŨÐãªé»ÃÐ⪹ì¨Ò¡ÊÁºÑµÔ¢Í§¤ÇÒÁà» ¹áÁèàËÅ硢ͧÇÑʴغҧª¹Ô´ ·Õ àÁ× ÍÍÂÙèã¹Ê¶Ò¹ÐÍÔ ÁµÑÇ (saturated) áÅéÇ ¨Ð·ÓãËéÊÀÒ¾¤ÇÒÁà» ¹áÁèàËÅç¡ã¹Ê× ÍºÑ¹·Ö¡ÁÕ·ÔÈ·Ò§ªÕ ä»ã¹·Ôȷҧ㴷ÔÈ·Ò§Ë¹Ö § ËÃ×Í ã¹·ÔÈ·Ò§µÃ§¡Ñ¹¢éÒÁ «Ö § ÅѡɳСÒúѹ·Ö¡ ¢éÍÁÙŠẺ¹Õ ¨ÐàËÁÒÐÊÓËÃѺ ¡ÒÃà¡çº ¢éÍÁÙÅ ´Ô¨Ô·ÑÅ ·Õ ÁÕ 2 ʶҹР¤×Í ºÔµ 1 áÅкԵ 0 ËÃ×Í·Õ àÃÕ¡¡Ñ¹ÇèÒ ¢éÍÁÙÅ亹ÒÃÕ (binary data) à¾ÃÒÐ©Ð¹Ñ ¹ 1
2
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
write signal
read-back signal write head
read head
medium disk motion
write current
read voltage
ÃÙ»·Õ 1.1: ËÅÑ¡¡ÒÃ¾× ¹°Ò¹¢Í§¡Òúѹ·Ö¡ÃкºáÁèàËÅç¡
ÇÑÊ´Ø àËÅèÒ¹Õ ¨Ö§ ¶Ù¡ ¹ÓÁÒ·Óà» ¹ Ê× Í ºÑ¹·Ö¡ à¾× Í à¡çº ¢éÍÁÙŠ亹ÒÃÕ à¹× ͧ¨Ò¡¢éÍÁÙÅ ã¹» ¨¨ØºÑ¹ ÊèǹÁÒ¡¨Ð ÍÂÙè ã¹ÃÙ» ¢Í§¢éÍÁÙÅ ´Ô¨Ô·ÑÅ àªè¹ ¢éÍÁÙÅ ã¹à¤Ã× Í§¤ÍÁ¾ÔÇàµÍÃì áÅТéÍÁÙÅ ·Õ ÃѺÊè§ ¼èÒ¹à¤Ã×Í¢èÒÂÍÔ¹ à·ÍÃì à¹çµ à» ¹µé¹ ¹Í¡¨Ò¡¹Õ ¢éÍÁÙÅá͹ÐÅçÍ¡¡çÊÒÁÒö·Õ ¨Ð¶Ù¡á»Å§ãËéÍÂÙèã¹ÃÙ»¢Í§¢éÍÁÙÅ´Ô¨Ô·ÑÅä´éà¾× ÍãËé §èÒµèÍ ¡ÒèѴ à¡çº ¢éÍÁÙÅ â´Â¼èÒ¹¢Ñ ¹µÍ¹¡ÒÃ¡Å Ó ÃËÑÊ ¾ÑÅÊì (PCM: pulse code modulation) [2] à¾ÃÒÐ©Ð¹Ñ ¹ ¡Òúѹ·Ö¡ÃкºáÁèàËÅç¡áºº´Ô¨Ô·ÑŨ֧àËÁÒÐÊÁ¡Ñº¡ÒÃà¡çº¢éÍÁÙÅã¹» ¨¨ØºÑ¹ ã¹» ¨¨ØºÑ¹¹Õ ¤ÇÒÁµéͧ¡ÒÃà¹× Í·Õ ã¹¡ÒèѴà¡çº¢éÍÁÙŢͧÍØ»¡Ã³ìÍÔàÅç¡·Ã͹ԡÊìµèÒ§æ ä´éá¡è ¤ÍÁ ¾ÔÇàµÍÃì, â·ÃÈѾ·ìà¤Å× Í¹·Õ , à¤Ã× Í§àÅè¹ à¾Å§áºº¾¡¾Ò, áÅСÅéͧ¶èÒÂÃÙ» ´Ô¨Ô·ÑÅ à» ¹µé¹ ÁÕ ÁÒ¡¢Ö ¹ àÃ× ÍÂæ à·¤â¹âÅÂÕ ¡Òúѹ·Ö¡ ÃкºáÁèàËÅç¡ áºº´Ô¨Ô·ÑÅ ¶×Í ä´é ÇèÒ à» ¹ ÇÔ¸Õ¡ÒÃËÅÑ¡ ·Õ ãªé 㹡ÒèѴ à¡çº ¢éÍÁÙŢͧ§Ò¹»ÃÐÂØ¡µì (application) µèÒ§æ ÃÇÁ件֧ á¼è¹ºÑ¹·Ö¡áÁèàËÅç¡ (magnetic oppy disk), ᶺºÑ¹·Ö¡ áÁèàËÅç¡ (magnetic tape), ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì, á¼è¹ «Õ ´Õ (CD: compact disc), áÅÐá¼è¹ ´ÕÇÕ´Õ (DVD: digital versatile disc) à» ¹µé¹ ÍÂèÒ§äáçµÒÁ ·Ø¡§Ò¹»ÃÐÂØ¡µì¨ÐµÑ §ÍÂÙ躹¾× ¹°Ò¹¢Í§ ËÅÑ¡¡Ò÷ӧҹà´ÕÂǡѹ«Ö §à¡Õ ÂÇ¢éͧ¡Ñº ËÑÇÍèÒ¹ (read head), ËÑÇà¢Õ¹ (write head), áÅÐÊ× ÍºÑ¹·Ö¡ áÁèàËÅç¡ (magnetic media) ´Ñ§áÊ´§ã¹ÃÙ»·Õ 1.1 àÁ× Í ËÑÇÍèÒ¹áÅÐËÑÇà¢Õ¹ẺÍÔ¹´Ñ¡·Õ¿ (inductive head) ¨Ð·ÓÁÒ¨Ò¡ÊÒÃáÁèàËÅç¡ÃÙ»à¡×Í¡ÁéÒ·Õ ÁÕ¤èÒÊÀҾźÅéÒ§áÁèàËÅç¡ (coercivity) µ Ó áÅФèÒÊÀÒ¾
1.2.
à຺¨ÓÅͧ¢Í§Ãкº¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì
3
ãËé«ÖÁ¼èÒ¹ä´é (permeability) ÊÙ§ [1, 3] â´Â¨ÐÁÕ¢´ÅÇ´¾Ñ¹ÍÂÙèÃͺæ áÅÐÊ× ÍºÑ¹·Ö¡¨Ð·ÓÁÒ¨Ò¡ÊÒà áÁèàËÅç¡·Õ ÁÕ¤èÒÊÀҾźÅéÒ§áÁèàËÅç¡ÊÙ§ ã¹Ë¹Ñ§Ê×Í àÅèÁ ¹Õ ¨Ð¡ÅèÒǶ֧ ੾ÒÐà·¤â¹âÅÂÕ ¡Òúѹ·Ö¡ ¢éÍÁÙÅ 2 Ẻ ¤×Í ¡Òúѹ·Ö¡ Ẻá¹Ç ¹Í¹ (longitudinal recording) áÅСÒúѹ·Ö¡áººá¹ÇµÑ § (perpendicular recording) â´Â·Õ ෤⹠âÅÂÕ ¡Òúѹ·Ö¡ Ẻá¹Ç¹Í¹à» ¹ à·¤â¹âÅÂÕ ·Õ ãªé 㹡Òúѹ·Ö¡ ¢éÍÁÙÅ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì µÑ §áµè Í´Õµ ¨¹ ¶Ö§ » ¨¨ØºÑ¹ ¹Ñ ¹¤×Í ÊÀÒ¾¤ÇÒÁà» ¹ áÁèàËÅç¡ ¢Í§Ê× Í ºÑ¹·Ö¡ ¨Ð¢¹Ò¹¡Ñº ÃйҺ¢Í§¨Ò¹ºÑ¹·Ö¡ áÁèàËÅç¡ (magnetic disk) ´Ñ§·Õ áÊ´§ã¹ÃÙ»·Õ 1.1 ã¹¢³Ð·Õ à·¤â¹âÅÂÕ¡Òúѹ·Ö¡áººá¹ÇµÑ §ä´éàÃÔ Á·Õ ¨Ð¹ÓÁÒãªé ÊÓËÃѺ ¡Òúѹ·Ö¡ ¢éÍÁÙÅ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ã¹» ¨¨ØºÑ¹ â´ÂÊÀÒ¾¤ÇÒÁà» ¹ áÁèàËÅç¡ ¢Í§Ê× Í ºÑ¹·Ö¡ ¨Ð µÑ §©Ò¡¡ÑºÃйҺ¢Í§¨Ò¹ºÑ¹·Ö¡áÁèàËÅç¡ «Ö §ã¹» ¨¨ØºÑ¹¹Õ §Ò¹ÇԨѷҧ´éҹ෤â¹âÅÂÕ¡Òúѹ·Ö¡¢éÍÁÙŠẺá¹ÇµÑ §ä´é´Óà¹Ô¹ä»ÍÂèÒ§ÃÇ´àÃçÇ à¾ÃÒÐÇèÒ à·¤â¹âÅÂÕ¡Òúѹ·Ö¡áººá¹Ç¹Í¹à¢éÒã¡Åé ¢Õ´¨Ó¡Ñ´ «Ùà»ÍÃì¾ÒÃÒáÁ¡à¹µÔ¡ (superparamagnetic limit) [1, 3, 4, 10] ·ÓãËéäÁèÊÒÁÒöà¾Ô Á¤ÇÒÁ¨Ø¢éÍÁÙÅ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ä´é ÁÒ¡¡ÇèÒ 1 à·ÃÐ亵ì (TB: terabyte) ¹Í¡¨Ò¡¹Õ à·¤â¹âÅÂÕ ¡Òúѹ·Ö¡ ¢éÍÁÙŠẺá¹ÇµÑ § ÊÒÁÒöªèÇÂà¾Ô Á ¢¹Ò´¤ÇÒÁ¨Ø ¢éÍÁÙÅ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ä´é ËÅÒÂÊÔº à·èÒ àÁ× Í à·Õº¡Ñº ¡Òà ãªéà·¤â¹âÅÂÕ¡Òúѹ·Ö¡¢éÍÁÙÅẺá¹Ç¹Í¹ [3, 5]
1.2
à຺¨ÓÅͧ¢Í§Ãкº¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì
Ãкº¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅ (digital data storage system) ã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ìÊÒÁÒö·Õ ¨Ð¨ÓÅͧ໠¹ á¼¹ÀÒ¾·Ñ Çä»ä´é µÒÁÃÙ»·Õ 1.2 àÁ× Í ºÔµ¢èÒÇÊÒà (message bits) ¨Ð¶Ù¡·Ó¡ÒÃà¢éÒÃËÑÊâ´Â ǧ¨Ãà¢éÒ ÃËÑÊá¡é䢢éͼԴ¾ÅÒ´ (error correction code (ECC) encoder) â´Â·Õ ÃËÑÊ RS (Reed Solomon code) [7, 8] à» ¹ ÃËÑÊ ·Õ ¹ÔÂÁ¹ÓÁÒãªé 㹡ÒÃà¢éÒ ÃËÑÊ á¡é䢢éͼԴ¾ÅÒ´¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ã¹» ¨¨ØºÑ¹ ¨Ò¡¹Ñ ¹ ¢éÍÁÙÅ·Õ à¢éÒÃËÑÊáÅéÇ¡ç¨Ð¶Ù¡·Ó¡ÒÃà¢éÒÃËÑÊÍÕ¡¤ÃÑ §Ë¹Ö §´éÇ ǧ¨Ãà¢éÒÃËÑÊÁÍ´ÙàŪѹ (modu lation encoder) à¾× Í·Ó˹éÒ·Õ ã¹¡ÒûÃѺ¤Ø³ÊÁºÑµÔ¢Í§¢éÍÁÙÅãËéàËÁÒÐÊÁ¡ÑºªèͧÊÑÒ³¢Í§ÎÒÃì´ ´ÔÊ¡ìä´Ã¿ì àªè¹ ·ÓãËéÅӴѺ¢éÍÁÙÅ (data sequence) ÁÕÃٻẺµÒÁ·Õ µéͧ¡Òà ËÃ×Í·ÓãËéÅӴѺ¢éÍÁÙÅäÁè ÁÕÊèǹ»ÃСͺ俿 Ò¡ÃÐáʵç (d.c. component) à» ¹µé¹ ÃËÑÊ·Õ ¹ÔÂÁãªéã¹Ç§¨Ãà¢éÒÃËÑÊÁÍ´ÙàŪѹ ¤×Í ÃËÑÊ RLL (run length limited code) [9] ¢éÍÁÙÅ àÍÒµì¾Øµ ·Õ ä´é ¨Ò¡Ç§¨Ãà¢éÒ ÃËÑÊ ÁÍ´ÙàŪѹ ¨Ð¶×Í ÇèÒ à» ¹ ¢éÍÁÙÅ ·Õ ¨Ð¶Ù¡ à¢Õ¹à¢éÒ ä»ã¹Ê× Í ºÑ¹·Ö¡ «Ö § ¨ÐàÃÕ¡¡Ñ¹ ÇèÒ ºÔµ ·Õ ¨Ð¶Ù¡ ºÑ¹·Ö¡ (recorded bit)
4
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
message bits
ECC encoder
modulation encoder
(e.g., RS codes)
(e.g., RLL codes)
write current waveform
recorded bits
modulator
write head/medium/read head assembly estimated message bits
ECC decoder
modulation decoder
reproduced bits
read channel read-back voltage waveform
A ÃÙ»·Õ 1.2: Ẻ¨ÓÅͧ·Ñ Ç仢ͧÃкº¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì [6]
ËÅѧ¨Ò¡¹Ñ ¹ ºÔµ ·Õ ¨Ð¶Ù¡ ºÑ¹·Ö¡ ¡ç ¨Ð¶Ù¡ Êè§ ä»Âѧ ǧ¨ÃÁÍ´ÙàÅàµÍÃì (modulator) à¾× Í á»Å§¢éÍÁÙÅ ºÔµ ãËé ÍÂÙè ã¹ÃÙ» ¤Å× ¹ ¡ÃÐáÊä¿¿ Ò à¢Õ¹ (write current waveform) ¨Ò¡¹Ñ ¹ ÃÙ» ¤Å× ¹ ¡ÃÐáÊä¿¿ Ò à¢Õ¹¡ç ¨Ð¶Ù¡» ͹ä»ÂѧËÑÇà¢Õ¹ à¾× Í·Ó¡ÒÃà¢Õ¹¢éÍÁÙÅŧä»ã¹Ê× ÍºÑ¹·Ö¡ ÊÓËÃѺ ¢Ñ ¹µÍ¹ã¹¡ÒÃÍèÒ¹¢éÍÁÙÅ ËÑÇÍèÒ¹¨Ð·Ó¡ÒÃÍèÒ¹¢éÍÁÙÅ ¨Ò¡Ê× Í ºÑ¹·Ö¡ àÁ× Í ËÑÇÍèÒ¹à¤Å× Í¹·Õ 1
ÁÒ¶Ö§ ºÃÔàdz·Õ ÁÕ ¡ÒÃà»ÅÕ Â¹á»Å§ÊÀÒ¾¤ÇÒÁà» ¹ áÁèàËÅç¡
(´Ù ÃÙ» ·Õ 1.1) ¨Ðä´é ¼ÅÅѾ¸ì ÍÍ¡ÁÒà» ¹
ÊÑÒ³ÃÙ»¤Å× ¹áç´Ñ¹ä¿¿ Ò ·Õ àÃÕ¡¡Ñ¹ÇèÒ ÊÑÒ³ read back ¨Ò¡¹Ñ ¹ ÊÑÒ³ read back ¡ç¨Ð ¶Ù¡Êè§à¢éÒä»·Ó¡ÒûÃÐÁÇżÅ㹪èͧÊÑÒ³ÍèÒ¹ (read channel) «Ö §»ÃСͺ仴éÇÂÊèǹ»ÃСͺ µèÒ§æ ä´éá¡è ǧ¨Ã¡Ãͧ¼èÒ¹µ Ó (LPF: low pass lter), ǧ¨ÃªÑ¡ µÑÇÍÂèÒ§ (sampler ËÃ×Í analog to digital converter), ÍÕ¤ÇÍäÅà«ÍÃì (equalizer), áÅÐǧ¨ÃµÃǨËÒ (detector) à» ¹µé¹ â´Â¢éÍÁÙÅ àÍÒµì¾Øµ ·Õ ä´é ¡ç ¨Ð¶Ù¡ ·Ó¡ÒöʹÃËÑÊ ´éÇ ǧ¨Ã¶Í´ÃËÑÊ ÁÍ´ÙàŪѹ (modulation decoder) áÅÐǧ¨Ã ¶Í´ÃËÑÊá¡é䢢éͼԴ¾ÅÒ´ (ECC decoder) à¾× ÍËÒ¤èÒ»ÃÐÁÒ³¢Í§ºÔµ¢èÒÇÊÒÃ·Õ Êè§ÁÒ¨Ò¡µé¹·Ò§ 1
ã¹·Ò§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì ºÃÔàdz·Õ ÁÕ¡ÒÃà»ÅÕ Â¹á»Å§ÊÀÒ¾¤ÇÒÁà» ¹áÁèàËÅç¡ ¨Ð¶Ù¡á·¹´éÇ¢éÍÁÙźԵ 1 áÅкÃÔàdz
·Õ ÁÕäÁè¡ÒÃà»ÅÕ Â¹á»Å§ÊÀÒ¾¤ÇÒÁà» ¹áÁèàËÅç¡ ¨Ð¶Ù¡á·¹´éÇ¢éÍÁÙźԵ 0 â´Â·Õ ÃٻẺ¢éÍÁÙÅÅÑ¡É³Ð¹Õ ¨ÐàÃÕ¡¡Ñ¹ÇèÒ ÃٻẺ NRZI (non return to zero interleaved) àÁ× Í ¢éÍÁÙÅ ºÔµ 1 ËÁÒ¶֧ ÁÕ ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð (transition) áÅÐ ¢éÍÁÙźԵ 0 ËÁÒ¶֧ äÁèÁÕ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
1.3.
¡Ãкǹ¡ÒÃà¢Õ¹
1.3
5
¡Ãкǹ¡ÒÃà¢Õ¹
ã¹ÃÐËÇèÒ§¡Ãкǹ¡ÒÃà¢Õ¹¢éÍÁÙÅ [10] ¢éÍÁÙźԵ¨Ð¶Ù¡á»Å§ãËéÍÂÙèã¹ÃÙ»¤Å× ¹¡ÃÐáÊä¿¿ ÒÃÙ»ÊÕ àËÅÕ ÂÁ (rectangular current waveform) ·Õ àÃÕ¡¡Ñ¹ ÇèÒ ¡ÃÐáÊä¿¿ Ò à¢Õ¹ (write current) (´Ù ÃÙ» ·Õ 1.1) â´Âǧ¨ÃÁÍ´ÙàÅàµÍÃì (modulator) [1, 4] ¨Ò¡¹Ñ ¹ ¡ÃÐáÊä¿¿ Ò à¢Õ¹¨Ð¶Ù¡ » ͹ä»Âѧ ¢´ÅÇ´¢Í§ËÑÇ à¢Õ¹ (write head) ·ÓãËéà¡Ô´à» ¹Ê¹ÒÁà¢Õ¹áÁèàËÅç¡ (magnetic write eld) ºÃÔàdzªèͧÇèÒ§ (gap) ÃÐËÇèÒ§ Ê× Í ºÑ¹·Ö¡ ¡Ñº ËÑÇ à¢Õ¹ â´Â ·Ñ Çä» Ê¹ÒÁ à¢Õ¹ áÁèàËÅç¡ ¨Ð µéͧ ÁÕ ¢¹Ò´ ËÃ×Í ¤ÇÒÁ à¢éÁ ÁÒ¡ ¡ÇèÒ ÊÀҾźÅéÒ§áÁèàËÅ硢ͧÊ× Í ºÑ¹·Ö¡ à¾× Í·Õ ¨Ðä´é ÊÒÁÒö·ÓãËé Ê× Í ºÑ¹·Ö¡ ³ ºÃÔàdz¹Ñ ¹ ÁÕ ÊÀÒ¾¤ÇÒÁ à» è¹ áÁèàËÅç¡ µÒÁ·ÔÈ·Ò§¢Í§Ê¹ÒÁà¢Õ¹áÁèàËÅç¡ ·Õ » ͹à¢éÒ ä» ¹Í¡¨Ò¡¹Õ ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹ÐÊÀÒ¾ ¤ÇÒÁà» ¹ áÁèàËÅç¡ (magnetization transition) ¢Í§Ê× Í ºÑ¹·Ö¡ ÊÒÁÒö·Óä´é â´Â¡ÒÃà»ÅÕ Â¹á»Å§ ·ÔÈ·Ò§¢Í§Ê¹ÒÁáÁèàËÅç¡ÊÓËÃѺà¢Õ¹ (ËÃ×Í·ÔÈ·Ò§¢Í§¡ÃÐáÊä¿¿ Òà¢Õ¹) à¾× ÍãËéÊÍ´¤Åéͧ¡Ñº¡Òà à¢Õ¹¢éÍÁÙźԵ 0 áÅкԵ 1 ã¹·Ò§»¯ÔºÑµÔ Ãкº¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì¨Ðãªé ¡Òúѹ·Ö¡áººäº¹ÒÃÕ (bina 2
ry recording) ¹Ñ ¹¤×Í ÊÀÒ¾¤ÇÒÁà» ¹ áÁèàËÅç¡ ·Õ ÍÂÙè ã¹Ê× Í ºÑ¹·Ö¡ ¨ÐÁÕ à¾Õ§ 2 ·ÔÈ·Ò§à·èÒ¹Ñ ¹
ËÃ×Í
¡ÅèÒÇÍÕ¡ ¹ÑÂ Ë¹Ö § ¤×Í ÃкºÊÒÁÒöºÑ¹·Ö¡ ¢éÍÁÙÅ ä´é à¾Õ§ 2 ÃдѺ (ËÃ×Í 2 ¤èÒ) à·èÒ¹Ñ ¹ «Ö § µèÒ§¨Ò¡ ¡Òúѹ·Ö¡¢éÍÁÙŢͧ´ÕÇÕ´Õ (DVD) ·Õ ÊÒÁÒöºÑ¹·Ö¡¢éÍÁÙÅä´éËÅÒÂæ ÃдѺ ·Ñ §¹Õ à» ¹à¾ÃÒÐÇèÒ â´Â»¡µÔ ¡Ãкǹ¡ÒÃà¢Õ¹¢éÍÁÙÅÁÕ¤ÇÒÁäÁèà» ¹àªÔ§àÊé¹ (nonlinearity) ÍÂÙè¾ÍÊÁ¤Çà ´Ñ§¹Ñ ¹ ¶éÒ·Ó¡Òúѹ·Ö¡ ¢éÍÁÙÅÁÒ¡¡ÇèÒ 2 ÃдѺŧä»ã¹Ê× ÍºÑ¹·Ö¡¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì ¼Å¡Ãзº·Õ à¡Ô´¨Ò¡¤ÇÒÁäÁèà» ¹àªÔ§àÊé¹ ¡ç¨ÐÂÔ §ÁÕ¤ÇÒÁÃعáçÁÒ¡¢Ö ¹ «Ö §¨ÐÊ觼ŷÓãËé»ÃÐÊÔ·¸ÔÀÒ¾ÃÇÁ¢Í§ÃкºáÂèŧÁÒ¡ [4]
1.4
¡Ãкǹ¡ÒÃÍèÒ¹
ã¹ÃÐËÇèÒ§¡Ãкǹ¡ÒÃÍèÒ¹¢éÍÁÙÅ [10] ËÑÇÍèÒ¹¨Ð·Ó¡ÒõÃǨ¨Ñº¡ÒÃà»ÅÕ Â¹á»Å§¿ÅÑ¡«ìáÁèàËÅç¡ (mag netic ux) ³ µÓáË¹è§·Õ ÁÕ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹ÐÊÀÒ¾¤ÇÒÁà» ¹áÁèàËÅç¡ «Ö §à» ¹¼Å·ÓãËéà¡Ô´à» ¹ÊÑÒ³ ¾ÑÅÊì áç´Ñ¹ä¿¿ Ò àË¹Õ ÂǹÓã¹¢´ÅÇ´ µÒÁ¡®¢Í§¿ÒÃÒà´Âì (Faraday s law) ÊÓËÃѺ ºÃÔàdz·Õ ÁÕ ¡Òà 2
¶éÒ ¾Ô¨ÒóҷÔÈ·Ò§¢Í§Ê¹ÒÁáÁèàËÅç¡ ¨Ò¡¢Ñ Ç à˹×Í ä»¢Ñ Ç ãµé (ËÃ×Í ¨Ò¡¢Ñ ÇºÇ¡ä»¢Ñ Çź) ¡Òúѹ·Ö¡ Ẻá¹Ç¹Í¹¨ÐÁÕ
ÅѡɳÐà» ¹áºº¢ÇÒ仫éÒ ËÃ×Í«éÒÂ仢ÇÒ ã¹¢³Ð·Õ ¡Òúѹ·Ö¡áººá¹ÇµÑ §¨Ðà» ¹áººº¹Å§ÅèÒ§ ËÃ×ÍÅèÒ§¢Ö ¹º¹ à·èÒ¹Ñ ¹
6
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
à»ÅÕ Â¹Ê¶Ò¹ÐàÍ¡à·È (isolated transition) ËÑÇÍèÒ¹¨ÐãËé ÊÑÒ³¾ÑÅÊì áç´Ñ¹ä¿¿ Ò ·Õ àÃÕ¡¡Ñ¹ ÇèÒ ÊÑÒ³¾ÑÅÊì à»ÅÕ Â¹Ê¶Ò¹Ð (transition pulse)
g(t)
ËÃ×Í
−g(t)
â´Â¨Ð¢Ö ¹ ÍÂÙè ¡Ñº ·ÔÈ·Ò§¢Í§
ÊÀÒ¾¤ÇÒÁà» ¹áÁèàËÅç¡ã¹Ê× ÍºÑ¹·Ö¡ (´ÙÃÙ»·Õ 1.1) ÊÓËÃѺÃкº¡Òúѹ·Ö¡áººá¹Ç¹Í¹ ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹Ð (ËÃ×Í ÊÑÒ³¾ÑÅÊì Lorent zian) ÊÒÁÒöà¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§ÊÁ¡Ò÷ҧ¤³ÔµÈÒʵÃìä´é ¤×Í [4]
g(t) =
³ 1+
PW50
àÁ× Í
¤×Í ¤ÇÒÁ¡ÇéÒ§¢Í§ÊÑÒ³¾ÑÅÊì
g(t)
1 2t PW50
´2
(1.1)
ÇÑ´ ³ µÓáË¹è§ ·Õ ÊÑÒ³¾ÑÅÊì ÁÕ ¤ÇÒÁÊÙ§ à» ¹
¤ÃÖ §Ë¹Ö § ¢Í§¤ÇÒÁÊÙ§ ÊÙ§ÊØ´ áÅÐÊÓËÃѺ ¡Òúѹ·Ö¡ Ẻá¹ÇµÑ § ÊÑÒ³¾ÑÅÊì à»ÅÕ Â¹Ê¶Ò¹Ð¨ÐÁÕ ÃÙ» ÊÁ¡Òà ¤×Í [11]
ln(·)
àÁ× Í
¤×Í ÅÍ¡ÒÃÔ·ÖÁ¸ÃÃÁªÒµÔ (natural logarithm),
tion) «Ö §¹ÔÂÒÁâ´Â ¢Í§
g(t)
à √ ! 2t ln 2 g(t) = erf PW50
erf(x) =
√2 π
Rx 0
e
−t2
erf(·)
dt, áÅÐ PW50
(1.2)
¤×Í ¿ §¡ìªÑ¹¢éͼԴ¾ÅÒ´ (error func
¤×Í ¤ÇÒÁ¡ÇéÒ§¢Í§¾ÑÅÊì
g 0 (t) ËÃ×Í Í¹Ø¾Ñ¹¸ì
ÇÑ´ ³ µÓáË¹è§·Õ ÊÑÒ³¾ÑÅÊìÁÕ¤ÇÒÁÊ٧໠¹¤ÃÖ §Ë¹Ö §¢Í§¤ÇÒÁÊÙ§ÊÙ§ÊØ´
ã¹ Ãкº ¡Òà ºÑ¹·Ö¡ ¢éÍÁÙÅ ¢Í§ ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ¤ÇÒÁ ˹Òá¹è¹ ¢Í§ ¡Òà ºÑ¹·Ö¡ Ẻ ¹ÍÃì ÁÍ ÅäÅ«ì (ND: normalized recording density) ËÃ×Í ¤ÇÒÁ˹Òá¹è¹¢Í§¡Òúѹ·Ö¡¢éÍÁÙÅ [4] ¨Ð¹ÔÂÒÁâ´Â
ND = àÁ× Í
T
(1.3)
¤×Í ¤ÒºàÇÅҢͧ¢éÍÁÙÅË¹Ö §ºÔµ ËÃ×Í·Õ àÃÕ¡¡Ñ¹ÇèÒ ºÔµà«ÅÅì (bit cell) «Ö §¨Ðà» ¹µÑǺ觺͡
ÇèÒ ºÃÔàdz ¤èÒ
PW50 T
PW50
PW50
ÊÒÁÒö·Õ ¨Ð¨Ñ´à¡çº¢éÍÁÙÅä´é¨Ó¹Ç¹¡Õ ºÔµ ´Ñ§¹Ñ ¹ ¶éÒ¡Ó˹´ãËé
T
à» ¹¤èÒ¤§·Õ àÁ× Í
ËÃ×Í ND à¾Ô Á ¢Ö ¹ ¡ç ËÁÒ¤ÇÒÁÇèÒ ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ÊÒÁÒö¨Ø ¢éÍÁÙÅ ä´é ÁÒ¡¢Ö ¹ ÃÙ» ·Õ 1.3
áÊ´§¼ÅµÍºÊ¹Í§¡ÒÃà»ÅÕ Â¹Ê¶Ò¹ÐÊÓËÃѺ¡Òúѹ·Ö¡áººá¹Ç¹Í¹áÅÐẺá¹ÇµÑ § ³ ÃдѺ ND µèÒ§æ ¨ÐàËç¹ä´éÇèÒ ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹Ð¢Í§·Ñ § 2 Ãкº¨Ð¤Ãͺ¤ÅØÁªèǧàÇÅÒËÅÒÂæ ºÔµà«ÅÅì â´Â੾ÒÐÍÂèÒ§ÂÔ § àÁ× Í ND ÁÕ ¤èÒ à¾Ô Á ¢Ö ¹ ËÃ×Í ÍÒ¨¨Ð¡ÅèÒÇä´é ÇèÒ ¡ÒÃá·Ã¡ÊÍ´ÃÐËÇèÒ§ÊÑÅѡɳì (ISI: intersymbol interference) ã¹ÊÑÒ³ read back ¨ÐÁÕ ¤ÇÒÁÃعáçÁÒ¡¢Ö ¹ àÁ× Í ND ÁÕ ¤èÒ
1.4.
¡Ãкǹ¡ÒÃÍèÒ¹
7
ND = 2
1
ND = 2.5 ND = 3
Amplitude
0.8
0.6
0.4
0.2
0 −5
0
5
(a) t/T 1 0.8
ND = 2 ND = 2.5 ND = 3
0.6
Amplitude
0.4 0.2 0 −0.2 −0.4 −0.6 −0.8 −1 −5
0
5
(b) t/T
ÃÙ»·Õ 1.3: ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹Ð ÊÓËÃѺ¡Òúѹ·Ö¡ (a) Ẻá¹Ç¹Í¹ áÅÐ (b) Ẻá¹ÇµÑ §
à¾Ô Á¢Ö ¹ à¹× ͧ¨Ò¡ âÍ¡ÒÊ·Õ ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹Ð·Õ ÍÂÙèã¡Åé¡Ñ¹¨ÐÁÒ«é͹àËÅ× ÍÁ (overlap) ¡Ñ¹ÁÕ ¤ÇÒÁà» ¹ä»ä´éÊÙ§ ã¹¡Ã³Õ·Õ ËÑÇÍèÒ¹à¤Å× Í¹·Õ ÁÒ¶Ö§ºÃÔàdz·Õ ÁÕµÓá˹觡ÒÃà»ÅÕ Â¹Ê¶Ò¹ÐµÔ´¡Ñ¹ 2 ¤ÃÑ § ÊÑÒ³¾ÑÅÊì
8
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ÊØ·¸Ô ·Õ ä´é ¨ÐàÃÕ¡¡Ñ¹ ÇèÒ ÊÑÒ³¾ÑÅÊìä´ºÔµ (dibit pulse) ËÃ×Í ¼ÅµÍºÊ¹Í§ä´ºÔµ (dibit re sponse) [4] «Ö §ÁÕ¤èÒà·èҡѺ
m(t) = g(t) − g(t − T )
(1.4)
´Ñ§áÊ´§ã¹ÃÙ»·Õ 1.4 ¶éÒãªé¡ÒÃá»Å§¿ÙàÃÕÂÃì·Õ µèÍà¹× ͧ·Ò§àÇÅÒ (continuous time Fourier transform) [12] ¡ÑºÊÑÒ³
m(t)
m(t)
ÊÓËÃѺ Ãкº¡Òúѹ·Ö¡
M (Ω) = exp{−π|Ω|ND} (1 − exp{−j2πΩ})
(1.5)
¨Ðä´é ÇèÒ ¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ (frequency response) ¢Í§
Ẻá¹Ç¹Í¹ ¤×Í
àÁ× Í
exp{·}
m(t)
¤×Í ¿ §¡ìªÑ¹àÅ¢ªÕ ¡ÓÅѧ (exponential function) ã¹¢³Ð·Õ ¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ ¢Í§
ÊÓËÃѺÃкº¡Òúѹ·Ö¡áººá¹ÇµÑ § ¤×Í
½ 2 2 ¾ π Ω ND2 T exp − (1 − exp{−j2πΩ}) M (Ω) = jπΩ ln(16) àÁ× Í
Ω = fT
(1.6)
f ¤×Í ¤ÇÒÁ¶Õ ÁÕ˹èÇÂà» ¹ √ j = −1 ¤×Í Ë¹èǨԹµÀÒ¾
¤×Í ¤ÇÒÁ¶Õ Ẻ¹ÍÃìÁÍÅäÅ«ì (normalized frequency),
àÎÔõ«ì (Hertz),
|x|
¤×Í ¤èÒÊÑÁºÙóì (absolute value) ¢Í§
x,
áÅÐ
(imadinary unit) ÃÙ»·Õ 1.5 áÊ´§¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ ¢Í§ÊÑÒ³¾ÑÅÊìä´ºÔµ ¨ÐàËç¹ä´éÇèÒ àÁ× Í ND à¾Ô Á ¢Ö ¹ ÃÙ»ÃèÒ§¢Í§¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§ÊÑÒ³¾ÑÅÊìä´ºÔµ ·Ñ § 2 Ẻ¨Ð¶Ù¡ ºÕº ãËé ÁÒ ÍÂÙè ³ ºÃÔàdz¤ÇÒÁ¶Õ µ Ó ¹Í¡¨Ò¡¹Õ ªèͧÊÑÒ³¢Í§¡Òúѹ·Ö¡áººá¹Ç¹Í¹¨ÐÁÕÊ໡µÃÑÁ¤èÒÈÙ¹Âì (spectral null) ³ µÓáË¹è§·Õ ¤ÇÒÁ¶Õ
f = 0 «Ö §ËÁÒ¶֧ äÁèÁÕÊèǹ»ÃСͺ俿 Ò¡ÃÐáʵç ã¹¢³Ð·Õ
ªèͧÊÑÒ³¢Í§¡Òúѹ·Ö¡áººá¹ÇµÑ §¨ÐÁÕÊèǹ»ÃСͺ俿 Ò¡ÃÐáʵç ËÁÒÂà˵Ø
3
ã¹Ë¹Ñ§Ê×Í àÅèÁ ¹Õ ¨Ðãªé â»Ãá¡ÃÁ SCILAB
[14] 㹡ÒÃÇÒ´ÃÙ» ¡ÃÒ¿¢Í§ÊÑÒ³µèÒ§æ
ÃÇÁ·Ñ § ¼Å¡Ò÷´Åͧ·Õ ä´é ¨Ò¡¡Ò÷ӡÒèÓÅͧ (simulation) Ãкº ¼ÙéÍèÒ¹¤ÇÃ·Õ ¨Ð¾ÂÒÂÒÁ·´Åͧ ÇÒ´ÃÙ»¡ÃÒ¿µèÒ§æ ã¹Ë¹Ñ§Ê×ÍàÅèÁ¹Õ à¾× Í·Õ ¨Ðä´éªèÇ·ÓãËéà¢éÒã¨ã¹º·àÃÕ¹ÁÒ¡ÂÔ §¢Ö ¹ 3
â»Ãá¡ÃÁ SCILAB à» ¹â»Ãá¡ÃÁ·Õ ÊÒÁÒö·Ó§Ò¹ä´éÍÂèÒ§ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ã¡Åéà¤Õ§¡Ñºâ»Ãá¡ÃÁ MATLAB [13] «Ö §
¤èÒÅÔ¢ÊÔ·¸Ô «Í¿µìáÇÃì ¢Í§â»Ãá¡ÃÁ MATLAB ÁÕ ÃÒ¤ÒᾧÁÒ¡ áµè â»Ãá¡ÃÁ SCILAB à» ¹ â»Ãá¡ÃÁ·Õ ãËé¿ÃÕ (freeware) ¼ÙéÍèÒ¹ÊÒÁÒö´ÒǹìâËÅ´µÑÇâ»Ãá¡ÃÁä´é¨Ò¡ http//www.scilab.org ËÃ×Í http://home.npru.ac.th/∼t3058/Scilab.html
1.5.
à຺¨ÓÅͧªèͧÊÑÒ³¡Òúѹ·Ö¡ÃкºààÁèàËÅç¡
0.5
9
ND = 2 ND = 2.5
0.4
ND = 3 0.3
Amplitude
0.2 0.1 0 −0.1 −0.2 −0.3 −0.4 −0.5 −5
0
5
(a) t/T 1 ND = 2
0.9
ND = 2.5 0.8
ND = 3
Amplitude
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 −5
0
5
(b) t/T
ÃÙ»·Õ 1.4: ¼ÅµÍºÊ¹Í§ä´ºÔµ ÊÓËÃѺ¡Òúѹ·Ö¡ (a) Ẻá¹Ç¹Í¹ áÅÐ (b) Ẻá¹ÇµÑ §
1.5
à຺¨ÓÅͧªèͧÊÑÒ³¡Òúѹ·Ö¡ÃкºààÁèàËÅç¡
â´Â·Ñ Çä» ªèͧÊÑÒ³¢Í§Ãкº¡Òúѹ·Ö¡ áÁèàËÅç¡ ÊÒÁÒö¨ÓÅͧä´é à» ¹ 2 Ẻ ¤×Í áºº¨ÓÅͧ ªèͧÊÑÒ³ àÊÁ×͹ ¨ÃÔ§ (realistic channel model) áÅРẺ¨ÓÅͧ ªèͧÊÑÒ³ ÍØ´Á¤µÔ (ideal
10
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ND = 2
1
Normalized magnitude
ND = 2.5 ND = 3 0.8
0.6
0.4
0.2
0
0
0.1
0.2
0.3
0.4
0.5
(a) Normalized frequency (fT)
ND = 2
1
ND = 2.5
Normalized magnitude
ND = 3 0.8
0.6
0.4
0.2
0
0
0.1
0.2
0.3
0.4
0.5
(b) Normalized frequency (fT)
ÃÙ»·Õ 1.5: ¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ ¢Í§ÊÑÒ³¾ÑÅÊìä´ºÔµ ÊÓËÃѺ¡Òúѹ·Ö¡ (a) Ẻá¹Ç¹Í¹ áÅÐ (b) Ẻá¹ÇµÑ §
1.5.
à຺¨ÓÅͧªèͧÊÑÒ³¡Òúѹ·Ö¡ÃкºààÁèàËÅç¡
11
n(t)
ak
1–D
bk
p(t)
g(t)
LPF
symbol detector
equalizer
â k
m(t) timing recovery target response H(D) ÃÙ»·Õ 1.6: Ẻ¨ÓÅͧªèͧÊÑÒ³àÊÁ×͹¨ÃÔ§
channel model) â´Â·Õ Ẻ¨ÓÅͧªèͧÊÑÒ³àÊÁ×͹¨ÃÔ§ [10] ¨ÐÁÕÅѡɳСÒ÷ӧҹã¡Åéà¤Õ§¡Ñº Ãкº¨ÃÔ§ à¹× ͧ¨Ò¡»ÃСͺ仴éÇ·ء Êèǹ»ÃСͺ·Õ ÊÓ¤Ñ ·Õ ÁÕ ÍÂÙè 㹠ʶһ µÂ¡ÃÃÁªèͧÊÑÒ³ ÍèÒ¹ (read channel architecture) [8, 10] ¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì ã¹¢³Ð·Õ Ẻ¨ÓÅͧªèͧÊÑÒ³ ÍØ´Á¤µÔÁÑ¡¨Ð¹ÔÂÁãªé㹡ÒÃÈÖ¡ÉÒ áÅÐÇÔà¤ÃÒÐËì¾× ¹°Ò¹¡Ò÷ӧҹ¢Í§Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³ ¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì à¹× ͧ¨Ò¡ à» ¹áºº¨ÓÅͧ·Õ äÁè«Ñº«é͹áÅЧèÒµèÍ¡Ò÷ӤÇÒÁà¢éÒã¨
1.5.1
Êèǹ
A
à຺¨ÓÅͧªèͧÊÑÒ³àÊÁ×͹¨ÃÔ§
ã¹ÃÙ»·Õ 1.2 ÊÒÁÒö·Õ ¨ÐáÊ´§ãËéÍÂÙèã¹ÃÙ»¢Í§áºº¨ÓÅͧ·Ò§¤³ÔµÈÒʵÃìä´é µÒÁÃÙ»·Õ 1.6
¡ÅèÒǤ×Í ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ Ẻ亹ÒÃÕ
ak ∈ {0, 1}
·Õ ÁÕ ¤ÒºàÇÅҢͧºÔµ (bit period) à·èÒ ¡Ñº
¨Ð¶Ù¡Ê觼èÒ¹ä»Âѧǧ¨ÃËÒ͹ؾѹ¸ìÍØ´Á¤µÔ (ideal di erentiator) ˹èǧàÇÅÒ
T
˹èÇ ·ÓãËéä´éà» ¹ÅӴѺ¢éÍÁÙÅà»ÅÕ Â¹Ê¶Ò¹Ð
1−D
àÁ× Í
D
¤×Í µÑÇ´Óà¹Ô¹¡ÒÃ
bk ∈ {−1, 0, 1} àÁ× Í bk = ±1 ËÁÒ¶֧
¡ÒÃà»ÅÕ Â¹Ê¶Ò¹ÐẺºÇ¡ (positive transition) ËÃ×Í áººÅº (negative transition) áÅÐ ËÁÒ¶֧ äÁèÁÕ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð ÅӴѺ¢éÍÁÙÅà»ÅÕ Â¹Ê¶Ò¹Ð á·¹´éǼŵͺʹͧ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð read back,
p(t),
g(t)
T
bk
bk = 0
¨Ð¶Ù¡Ê觼èÒ¹ä»ÂѧªèͧÊÑÒ³·Õ ¶Ù¡
áÅж١ ú¡Ç¹´éÇÂÊÑҳú¡Ç¹
n(t)
ÊÑÒ³
¨Ð¶Ù¡¡Ãͧ´éÇÂǧ¨Ã¡Ãͧ¼èÒ¹µ Ó (LPF) à¾× ͡ӨѴÊÑҳú¡Ç¹·Õ ÍÂÙè¹Í¡á¶º
¤ÇÒÁ¶Õ (out of band noise) ¨Ò¡¹Ñ ¹ ¡ç ¨Ð¶Ù¡ ·Ó¡Òêѡ µÑÇÍÂèÒ§ (sampling) ³ àÇÅÒ·Õ ¶Ù¡ ¤Çº¤ØÁ
12
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
â´ÂÃкºä·ÁÁÔ § ÃԤѿàÇÍÃÕ (timing recovery) ¢éÍÁÙÅ àÍÒµì¾Øµ ¢Í§Ç§¨ÃªÑ¡ µÑÇÍÂèÒ§¨Ð¶Ù¡ Êè§ ¼èÒ¹ä»Âѧ ÍÕ¤ÇÍäÅà«ÍÃì áÅÐǧ¨ÃµÃǨËÒÊÑÅѡɳì (symbol detector) à¾× ÍËÒÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ·Õ à» ¹ä»ä´é ÁÒ¡·Õ ÊØ´ (most likely input sequence) ¹Ñ ¹¤×Í ËÒ¤èÒ»ÃÐÁÒ³¢Í§
ak
ËÃ×Í
âk
ǧ¨ÃµÃǨËÒÊÑÅѡɳì·Õ ¹ÔÂÁãªéã¹Ãкº¡Òúѹ·Ö¡áÁèàËÅç¡ ¤×Í Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ (Viter bi detector) [15] ÍÂèÒ§äáçµÒÁ à¹× ͧ¨Ò¡ ¤ÇÒÁ«Ñº«é͹ (complexity) ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨Ðà¾Ô Á ¢Ö ¹ ẺàÅ¢ªÕ ¡ÓÅѧ µÒÁ¨Ó¹Ç¹Ë¹èǤÇÒÁ¨Ó¢Í§ªèͧÊÑÒ³ (channel memory) ´Ñ§¹Ñ ¹ ÍÕ¤ÇÍäÅà«ÍÃì¨Ö§ à» ¹ ÊÔ § ¨Óà» ¹ ·Õ ¨Ðµéͧ¶Ù¡ ¹ÓÁÒãªé §Ò¹ à¾× Í ·Ó˹éÒ·Õ ã¹¡ÒûÃѺ ÃÙ»ÃèÒ§¼ÅµÍºÊ¹Í§ 4
ÃÇÁ¢Í§·Ñ § ÃкºãËé à» ¹ ¼ÅµÍºÊ¹Í§·Õ µéͧ¡Òà ·Õ àÃÕ¡¡Ñ¹ ÇèÒ ¼ÅµÍºÊ¹Í§·ÒÃìà¡çµ sponse)
(target re
H(D) [4] áÅЪèÇ·ÓãËé¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔŴŧä´é (ÈÖ¡ÉÒÃÒÂÅÐàÍÕ´
à¾Ô ÁàµÔÁä´é㹺··Õ 3)
1.5.2
à຺¨ÓÅͧªèͧÊÑÒ³ÍØ´Á¤µÔ
¶éÒÊÁÁصÔãËé ÃкºÁÕ¡Ãкǹ¡ÒÃÍÕ¤ÇÍäÅ૪ѹẺÊÁºÙóì (perfect equalization) Ẻ¨ÓÅͧã¹ÃÙ» ·Õ 1.6 ¨ÐÊÒÁÒöŴÃÙ» ä´é à» ¹ Ẻ¨ÓÅͧªèͧÊÑÒ³ÍØ´Á¤µÔ ´Ñ§ áÊ´§ã¹ÃÙ» ·Õ 1.7 â´Â·Õ ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ Ẻ亹ÒÃÕ
ak
·Õ ÁÕ ¤ÒºàÇÅҢͧºÔµ à·èÒ ¡Ñº
ÊÑÒ³¾ÑÅÊì 乤ÇÔµÊì ÍØ´Á¤µÔ (ideal Nyquist pulse) ú¡Ç¹´éÇÂÊÑҳú¡Ç¹
n(t)
T
q(t)
¨Ð¶Ù¡ ¡Å Ó ÊÑÒ³ (modulate) ¡Ñº =
sin(πt/T )/(πt/T )
ÊÑÒ³·Õ ǧ¨ÃÀÒ¤ÃѺ ä´é ÃѺ
p(t)
[16] áÅж١
¨Ð¶Ù¡ Êè§ ¼èÒ¹ä»Âѧ ǧ¨Ã¡Ãͧ
¼èÒ¹µ Ó à¾× ͡ӨѴÊÑҳú¡Ç¹·Õ ÍÂÙè¹Í¡á¶º¤ÇÒÁ¶Õ ¨Ò¡¹Ñ ¹ ¡ç¨Ð¶Ù¡·Ó¡ÒêѡµÑÇÍÂèÒ§ ³ àÇÅÒ·Õ ¶Ù¡¤Çº¤ØÁâ´ÂÃкºä·ÁÁÔ §ÃԤѿàÇÍÃÕ ¨Ò¡¹Ñ ¹ ¢éÍÁÙÅàÍÒµì¾Øµ¢Í§Ç§¨ÃªÑ¡µÑÇÍÂèÒ§¡ç¨Ð¶Ù¡Ê觼èÒ¹ä»Âѧ ǧ¨ÃµÃǨËÒÊÑÅѡɳì à¾× Í·Ó¡ÒÃËÒÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ·Õ à» ¹ä»ä´éÁÒ¡·Õ ÊØ´ ·ÒÃìà¡çµáºº¼ÅµÍºÊ¹Í§ºÒ§Êèǹ ËÃ×Í ·Õ àÃÕ¡ÇèÒ ·ÒÃìà¡çµáºº PR (partial response) [17] ·Õ à» ¹·Õ ÂÍÁÃѺ㹡Òúѹ·Ö¡áººá¹Ç¹Í¹ ¨ÐÁÕÃÙ»ÊÁ¡ÒÃà» ¹ [4]
H(D) = (1 − D)(1 + D)n 4
(1.7)
·ÒÃìà¡çµ (target) ã¹·Ò§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì ËÁÒ¶֧ ǧ¨Ã¡ÃͧẺàªÔ§àÊé¹·Õ Áըӹǹá·ç» (tap) ¹éÍ áÅж١Í͡Ẻ
ãËéÁռŵͺʹͧàªÔ§¤ÇÒÁ¶Õ àËÁ×͹¡Ñº¼ÅµÍºÊ¹Í§¢Í§ªèͧÊÑÒ³ãËéÁÒ¡·Õ ÊØ´ â´Â»ÃÒȨҡ¡ÒâÂÒÂÊÑҳú¡Ç¹
1.5.
à຺¨ÓÅͧªèͧÊÑÒ³¡Òúѹ·Ö¡ÃкºààÁèàËÅç¡
13
n(t)
ak
rk
H(D)
p(t)
q(t)
âk
symbol detector
LPF
timing recovery
ÃÙ»·Õ 1.7: Ẻ¨ÓÅͧªèͧÊÑÒ³ÍØ´Á¤µÔ
ã¹¢³Ð·Õ ·ÒÃìà¡çµáºº PR ·Õ à» ¹·Õ ÂÍÁÃѺ㹡Òúѹ·Ö¡áººá¹ÇµÑ § ¨ÐÍÂÙèã¹ÃÙ»¢Í§ [18]
H(D) = (1 + D)n
àÁ× Í
n
(1.8)
¤×Í àÅ¢¨Ó¹Ç¹àµçÁ ºÇ¡ ¨Ò¡ÊÁ¡Òà (1.8) ¨Ð¾ºÇèÒ Ãкº¡Òúѹ·Ö¡ Ẻá¹ÇµÑ § ¨ÐäÁè ÁÕ ¾¨¹ì
(1 − D)
·Ñ §¹Õ à» ¹ à¾ÃÒÐÇèÒ ªèͧÊÑÒ³¡Òúѹ·Ö¡ Ẻá¹ÇµÑ § ÁÕ Êèǹ»ÃСͺ俿 Ò¡ÃÐáʵç (´Ù
ÃÙ» ·Õ 1.5) ÃÙ» ·Õ 1.8 à»ÃÕºà·Õº¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§·ÒÃìà¡çµ ẺµèÒ§æ â´Â·Õ µÑÇàÅ¢·Õ ÍÂÙè ã¹à¤Ã× Í§ËÁÒÂǧàÅçº ÊÕ àËÅÕ ÂÁ
[1 0 − 1]
D
áÊ´§¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô áµèÅÐá·ç» ¢Í§·ÒÃìà¡çµ µÑÇÍÂèÒ§àªè¹ PR4
ËÁÒ¶֧ ·ÒÃìà¡çµ Ẻ PR4 (PR class IV) ·Õ ÁÕ ¿ §¡ìªÑ¹ ¶èÒÂâ͹ã¹â´àÁ¹
H(D) = 1 − D2 ã¹â´àÁ¹
[. . .]
¤×Í
ËÃ×Í EEPR2
[1 4 6 4 1]
D
[10] ¤×Í
ËÁÒ¶֧ ·ÒÃìà¡çµáºº EEPR2 ·Õ ÁÕ¿ §¡ìªÑ¹¶èÒÂâ͹
H(D) = 1 + 4D + 6D2 + 4D3 + D4
à» ¹µé¹
¨Ò¡ÃÙ» ·Õ 1.8 ¨Ð¾ºÇèÒ àÁ× Í ªèͧÊÑÒ³ÁÕ ¤èÒ ND à¾Ô Á ¢Ö ¹ ·ÒÃìà¡çµ·Õ ãªé ¡ç ¤ÇÃ·Õ ¨ÐÁÕ ¨Ó¹Ç¹á·ç» ÁÒ¡¢Ö ¹ (ÁÕ ¤èÒ
n
ÁÒ¡¢Ö ¹) à¾× Í·Õ ¨Ð·ÓãËé ¼ÅµÍºÊ¹Í§·ÒÃìà¡çµ ÁÕ ÅѡɳÐã¡Åéà¤Õ§¡Ñº ¼ÅµÍºÊ¹Í§
¢Í§ ªèͧÊÑÒ³ ãËé ÁÒ¡ ·Õ ÊØ´ «Ö § ¨Ð Êè§ ¼Å ·Ó ãËé ǧ¨Ã µÃǨËÒ ÇÕà·ÍÃìºÔ ·Ó§Ò¹ ä´é ÍÂèÒ§ ÁÕ »ÃÐÊÔ·¸ÔÀÒ¾ ÁÒ¡¢Ö ¹ (ÈÖ¡ÉÒÃÒÂÅÐàÍÕ´à¾Ô ÁàµÔÁ ä´é 㹺··Õ 3) ¹Í¡¨Ò¡¹Õ ¨Ò¡ÊÁ¡Òà (1.7) áÅÐ (1.8) ¨Ð¾º ÇèÒ ¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§·ÒÃìà¡çµ Ẻ PR ·Ø¡ µÑÇ ¨Ðà» ¹ àÅ¢¨Ó¹Ç¹àµçÁ ÍÂèÒ§äáçµÒÁ ¶éÒ ãªé ·ÒÃìà¡çµ ·Õ ÁÕ ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô à» ¹àÅ¢¨Ó¹Ç¹¨ÃÔ§ «Ö §¨ÐàÃÕ¡ÇèÒ ·ÒÃìà¡çµáºº GPR (generalized partial response target) »ÃÐÊÔ·¸ÔÀÒ¾ÃÇÁ¢Í§Ãкº·Õ ä´é¨ÐÁÕÁÒ¡¡ÇèÒ¡ÒÃãªé·ÒÃìà¡çµáºº PR [18, 19]
14
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
Normalized magnitude
1
0.8
0.6
0.4
Channel response (ND = 2) Channel response (ND = 2.5) PR4 [1 0 −1] (n = 1)
0.2
EPR4 [1 1 −1 −1] (n = 2) EEPR4 [1 2 0 −2 1] (n = 3) 0
0
0.1
0.2
0.3
0.4
0.5
(a) Normalized frequency (fT)
Channel response (ND = 2)
1
Channel response (ND = 2.5)
Normalized magnitude
PR2 [1 2 1] (n = 2) EPR2 [1 3 3 1] (n = 3)
0.8
EEPR2 [1 4 6 4 1] (n = 4) 0.6
0.4
0.2
0
0
0.1
0.2
0.3
0.4
0.5
(b) Normalized frequency (fT)
ÃÙ»·Õ 1.8: ¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§·ÒÃìà¡çµ ẺµèÒ§æ ÊÓËÃѺ ªèͧÊÑÒ³¡Òúѹ·Ö¡ (a) Ẻ á¹Ç¹Í¹ áÅÐ (b) Ẻá¹ÇµÑ §
1.6.
ÊÃØ»·éÒº·
1.6
15
ÊÃØ»·éÒº·
㹺·¹Õ ä´é¡ÅèÒǶ֧¾× ¹°Ò¹¢Í§¡Òúѹ·Ö¡ÃкºáÁèàËÅç¡áºº´Ô¨Ô·ÑÅ ÃÇÁ·Ñ §ËÅÑ¡¡Ò÷ӧҹ¢Í§¡Ãкǹ ¡ÒÃà¢Õ¹áÅСÒÃÍèÒ¹¢éÍÁÙÅã¹ÍØ»¡Ã³ìÎÒÃì´´ÔÊ¡ìä´Ã¿ì ¹Í¡¨Ò¡¹Õ Âѧä´é͸ԺÒ¶֧Ẻ¨ÓÅͧ¡Ò÷ӧҹ ¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì ·Ñ §áºº¨ÓÅͧªèͧÊÑÒ³àÊÁ×͹¨ÃÔ§ áÅÐẺ¨ÓÅͧªèͧÊÑÒ³ÍØ´Á¤µÔ â´Â·Õ Ẻ¨ÓÅͧ·Ñ §Êͧ¹Õ ¨Ð¶Ù¡¹ÓÁÒãªé㹡ÒèÓÅͧÃкº (system simulation) 㹺·µèÍæ ä» à¹× ͧ¨Ò¡ ¡ÒÃÇÔà¤ÃÒÐËìÃкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì¨ÐÍÒÈÑÂ¾× ¹°Ò¹·Ò§¤³Ôµ ÈÒʵÃì·Õ à¡Õ ÂÇ¢éͧ¡Ñº¡ÒûÃÐÁÇżÅÊÑÒ³ áÅСÒÃÊ× ÍÊÒôԨԷÑŤè͹¢éÒ§ÁÒ¡ ´Ñ§¹Ñ ¹ ¼ÙéÍèÒ¹¤ÇÃ·Õ ¨Ð ·º·Ç¹¤ÇÒÁÃÙéàËÅèÒ¹Õ ãËéà¢éÒ㨡è͹·Õ ¨ÐÈÖ¡ÉÒ˹ѧÊ×ÍàÅèÁ¹Õ «Ö §¤ÇÒÁÃÙé¾× ¹°Ò¹·Ò§¤³ÔµÈÒʵÃì´éÒ¹µèÒ§æ ·Õ à¡Õ ÂÇ¢éͧ¡Ñº¡ÒÃÇÔà¤ÃÒÐËìÃкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì ÊÒÁÒö·Õ ¨ÐÈÖ¡ÉÒä´é¨Ò¡ ˹ѧÊ×Í ¡ÒûÃÐÁÇżÅÊÑÒ³ÊÓËÃѺ¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅ àÅèÁ 1: ¾× ¹°Ò¹ªèͧÊÑÒ³ÍèÒ¹ à¢Õ¹ [10]
1.7
à຺½ ¡ËÑ´·éÒº·
1. ¨§Í¸ÔºÒÂËÅÑ¡¡Ò÷ӧҹ¢Í§¡Òúѹ·Ö¡ÃкºáÁèàËÅç¡
2. ¨§Í¸ÔºÒ¢éÍᵡµèÒ§ÃÐËÇèÒ§Ãкº¡Òúѹ·Ö¡áººá¹Ç¹Í¹áÅÐẺá¹ÇµÑ §
3. ¨§Í¸ÔºÒÂËÅÑ¡¡ÒÃàº× ͧµé¹¢Í§¡Ãкǹ¡ÒÃà¢Õ¹áÅСÒÃÍèÒ¹¢éÍÁÙÅã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì
4. ¨§Í¸ÔºÒÂËÅÑ¡¡Ò÷ӧҹ¢Í§áºº¨ÓÅͧªèͧÊÑÒ³àÊÁ×͹¨ÃÔ§
5. ¨§Í¸ÔºÒÂËÅÑ¡¡Ò÷ӧҹ¢Í§áºº¨ÓÅͧªèͧÊÑÒ³ÍØ´Á¤µÔ
6. ¨§Â¡µÑÇÍÂèÒ§·ÒÃìà¡çµáºº PR ÊÓËÃѺÃкº¡Òúѹ·Ö¡áººá¹Ç¹Í¹ÁÒÍÂèÒ§¹éÍ 4 ·ÒÃìà¡çµ
7. ¨§Â¡µÑÇÍÂèÒ§·ÒÃìà¡çµáºº PR ÊÓËÃѺÃкº¡Òúѹ·Ö¡áººá¹ÇµÑ §ÁÒÍÂèÒ§¹éÍ 4 ·ÒÃìà¡çµ
16
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
º··Õ 2
ä·ÁÁÔ §ÃԤѿàÇÍÃÕ
㹺·¹Õ ¨Ð͸ԺÒ¶֧ ¤ÇÒÁÊÓ¤Ñ ¢Í§ä·ÁÁÔ § ÃԤѿàÇÍÃÕ (timing recovery) ¾ÃéÍÁ·Ñ § ͸ԺÒÂËÅÑ¡¡Òà ·Ó§Ò¹¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä» (conventional timing recovery) «Ö §ÍÂÙ躹¾× ¹°Ò¹¢Í§ ǧ¨Ãà¿ÊÅçÍ¡ÅÙ» (PLL: phased lock loop) ¹Í¡¨Ò¡¹Õ Âѧ͸ԺÒ¶֧ ÇÔ¸Õ¡ÒÃÍ͡Ẻ¤èÒ¾ÒÃÒÁÔàµÍÃì ¢Í§Ç§¨Ãà¿ÊÅçÍ¡ÅÙ» â´ÂãªéẺ¨ÓÅͧ¡Ò÷ӧҹ¢Í§Ç§¨Ãà¿ÊÅçÍ¡ÅٻẺàªÔ§àÊé¹ (linearized PLL model) ÃÇÁ·Ñ §áÊ´§¼Å¡Ò÷´Åͧà¾× ÍãËéàË繶֧¤ÇÒÁÊӤѢͧÃкºä·ÁÁÔ §ÃԤѿàÇÍÃÕ
2.1
º·¹Ó
ã¹» ¨¨ØºÑ¹¹Õ ¤ÇÒÁà¨ÃÔ¡éÒÇ˹éÒ ·Ò§´éҹ෤â¹âÅÂÕ ¡ÒÃÊ× ÍÊÒà (communication) ä´é ¾Ñ²¹Òä»ÍÂèÒ§ ÃÇ´àÃçÇ â´Â੾ÒÐÍÂèÒ§ÂÔ §¡ÒÃÊ× ÍÊÒôԨԷÑÅ·Õ ãªéã¹ÃкºµèÒ§æ àªè¹ Ãкºâ·ÃÈѾ·ìà¤Å× Í¹·Õ áÅÐÃкº ¡ÒûÃÐÁÇżÅÊÑÒ³ÀÒÂã¹ÍØ»¡Ã³ìÎÒÃì´´ÔÊ¡ìä´Ã¿ì à» ¹µé¹ à¾ÃÒÐ©Ð¹Ñ ¹ ¤ÇÒÁµéͧ¡ÒÃ·Õ ¨Ðà¾Ô Á »ÃÐÊÔ·¸ÔÀÒ¾â´ÂÃÇÁ¢Í§Ãкº¨Ö§ à» ¹ ÊÔ § ·Õ ¨Óà» ¹ ÁÒ¡ à¾ÃÒÐÇèÒ ¨Ðä´é ÊÒÁÒöÃѺ áÅÐÊè§ ¢éÍÁÙÅ ä´é ÍÂèÒ§ ¹èÒàª× Ͷ×ÍÁÒ¡ÂÔ §¢Ö ¹ ã¹ÃкºÊ× ÍÊÒôԨԷÑÅ (digital communication system) ¡ÒÃÊè§ÊÑÒ³¨Ò¡µé¹·Ò§ä»Âѧ»ÅÒ ·Ò§ ÍØ»¡Ã³ì µé¹·Ò§¨Ð·Ó˹éÒ·Õ à»ÅÕ Â¹ÊÑÒ³´Ô¨Ô·ÑÅ (digital) ãËé à» ¹ ÊÑÒ³á͹ÐÅçÍ¡ (ana log) ¡è͹·Õ ¨Ð¶Ù¡Êè§ÍÍ¡ä»Âѧ»ÅÒ·ҧ àÁ× ÍÊÑÒ³á͹ÐÅçÍ¡ÁÒ¶Ö§·Õ ÍØ»¡Ã³ì»ÅÒ·ҧ ¡ç¨Ð¶Ù¡Êè§ 17
18
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ä»Âѧ ǧ¨ÃªÑ¡ µÑÇÍÂèÒ§ (sampler) à¾× Í ·Ó¡ÒÃá»Å§ÊÑÒ³á͹ÐÅçÍ¡ãËé ¡ÅѺ ä»à» ¹ ÊÑÒ³´Ô¨Ô·ÑÅ ã¹ÃÙ» ¢Í§ ¢éÍÁÙÅ ÇÔÂص (discrete data) ËÃ×Í ·Õ àÃÕ¡¡Ñ¹ ÇèÒ á«Áà» Å (sample) ¡Òêѡ µÑÇÍÂèÒ§ (sampling) ÊÑÒ³á͹ÐÅçÍ¡·Õ ¼Ô´¨Ñ§ËÇШзÓãËéà¡Ô´¼ÅàÊÕÂËÒÂÍÂèÒ§ÁÒ¡¡Ñº»ÃÐÊÔ·¸ÔÀÒ¾â´ÂÃÇÁ ¢Í§Ãкº ä·ÁÁÔ §ÃԤѿàÇÍÃըзÓ˹éÒ·Õ ã¹¡ÒÃà¢éҨѧËÇÐ (synchronize) ǧ¨ÃªÑ¡µÑÇÍÂèÒ§¡ÑºÊÑÒ³ á͹ÐÅçÍ¡·Õ ä´é ÃѺ à¾× ÍãËé ä´é ¢éÍÁÙÅ á«Áà» Å ·Õ ´Õ ·Õ ÊØ´ ÍÍ¡ÁÒ ¡è͹·Õ ¨Ð¹Ó¢éÍÁÙÅ á«Áà» Å àËÅèÒ¹Ñ ¹ ä»·Ó ¡ÒûÃÐÁÇÅ¼Å¢Ñ ¹ µèÍä» àªè¹ Êè§ µèÍä»Âѧ ÍÕ¤ÇÍäÅà«ÍÃì (equalizer) áÅÐǧ¨Ã¶Í´ÃËÑÊ (decoder) à» ¹µé¹ ´Ñ§¹Ñ ¹ä·ÁÁÔ §ÃԤѿàÇÍÃÕ¨Ö§¹Ñºä´éÇèÒà» ¹Í§¤ì»ÃСͺ·Õ ¤ÇÒÁÊÓ¤ÑÁÒ¡ÍÂèÒ§Ë¹Ö §ã¹ÃкºÊ× ÍÊÒà ´Ô¨Ô·ÑÅ à¾ÃÒÐÇèÒ ¶éÒ Ç§¨ÃªÑ¡ µÑÇÍÂèÒ§·Ó§Ò¹äÁè ´Õ ¡ç ¨ÐÊè§ ¼Å·ÓãËé ¢éÍÁÙÅ á«Áà» Å ·Õ ä´é äÁè ÁÕ ¤Ø³ÀÒ¾ËÃ×Í ÁÕ ¢éͼԴ¾ÅÒ´ÁÒ¡ áÅÐàÁ× Í Êè§ ¢éÍÁÙÅ á«Áà» Å àËÅèÒ¹Õ ä»Âѧ ÍÕ¤ÇÍäÅà«ÍÃì áÅÐǧ¨Ã¶Í´ÃËÑÊ ¡ç ¨Ð·ÓãËé ¼ÅÅѾ¸ì·Õ ä´éÁÕ¢éͼԴ¾ÅÒ´ÁÒ¡ ¹Ñ ¹¤×Í ÍѵÃÒ¢éͼԴ¾ÅÒ´ºÔµ (BER: bit error rate) ¢Í§Ãкº¨ÐÁÕ¤èÒ ÊÙ§ «Ö §à» ¹ÊÔ §·Õ ¤ÇèÐËÅÕ¡àÅÕ Â§ â´Â·Ñ Çä»ä·ÁÁÔ §ÃԤѿàÇÍÃըзӧҹÍÂÙ躹¾× ¹°Ò¹¢Í§Ç§¨Ãà¿ÊÅçÍ¡ÅÙ» (PLL) [4] «Ö §»ÃÐ¡Íºä» ´éÇ ǧ¨ÃµÃǨËÒ¢éͼԴ¾ÅÒ´·Ò§àÇÅÒ (TED: timing error detector), ǧ¨Ã¡ÃͧÅÙ» (loop lter), áÅÐǧ¨Ã VCO (voltage controlled oscillator) ã¹·Ò§»¯ÔºÑµÔ â¤Ã§ÊÃéÒ§¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕ¨ÐÁÕÍÂÙè 2 ÃٻẺ ¤×Í áºº¹ÔùÑ (deductive) áÅÐẺÍØ»¹Ñ (inductive) â´Â¨Ð¢Ö ¹ÍÂÙè¡ÑºÇèÒ ¢éÍÁÙÅ·Ò§àÇÅÒ ¶Ù¡ ´Ö§ ÍÍ¡ÁÒãªé ³ µÓáË¹è§ ¡è͹ËÃ×Í ËÅѧ ǧ¨ÃªÑ¡ µÑÇÍÂèÒ§ ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº¹ÔùÑ (deductive timing recovery) ¨Ð´Ö§¢éÍÁÙÅ·Ò§àÇÅÒ ·Õ àÃÕ¡¡Ñ¹ÇèÒ timing tone [16] ¨Ò¡ÊÑÒ³á͹ÐÅçÍ¡·Õ ´éÒ¹¢Òà¢éҢͧǧ¨ÃªÑ¡µÑÇÍÂèÒ§ µÒÁÃÙ»·Õ 2.1 â´Â·Õ ǧ¨Ã PLL ¶Ù¡¹Óãªéà¾× ÍÅ´¼Å¡Ãзº¢Í§ä·ÁÁÔ § 1
¨ÔµàµÍÃì
(timing jitter) ·Õ ὧÍÂÙèã¹ÊÑÒ³á͹ÐÅçÍ¡ ã¹¢³Ð·Õ ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺÍØ»¹Ñ (in
ductive timing recovery) ¨Ðãªéǧ¨Ã PLL Ẻ» ͹¡ÅѺ (feedback) à¾× Í·Ó˹éÒ·Õ ã¹¡Òô֧¢éÍÁÙÅ ·Ò§àÇÅÒ¨Ò¡ÊÑÒ³á͹ÐÅçÍ¡·Õ ´éÒ¹¢ÒÍÍ¡¢Í§Ç§¨ÃªÑ¡ µÑÇÍÂèÒ§ ´Ñ§·Õ áÊ´§ã¹ÃÙ» ·Õ 2.2 ¢éÍ´Õ ¢Í§ ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áººÍØ»¹Ñ ¤×Í ·Ø¡ Êèǹ»ÃСͺ¢Í§ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº¹Õ ÊÒÁÒö·Õ ¨Ð¶Ù¡ ÊÃéÒ§ãËé à» ¹áºº´Ô¨Ô·ÑÅä´é à¹× ͧ¨Ò¡ ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺÍØ»¹ÑÂà» ¹·Õ ¹ÔÂÁãªé§Ò¹¡Ñ¹ÁÒ¡ ´Ñ§¹Ñ ¹ ã¹Ë¹Ñ§Ê×Í àÅèÁ ¹Õ ¨ÐàÃÕ¡ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº¹Õ ÇèÒ ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº·Õ ãªé ¡Ñ¹ ·Ñ Çä» (conventional timing
1
ä·ÁÁÔ § ¨Ôµ àµÍÃì ¤×Í ÊÑҳú¡Ç¹ÀÒÂã¹Ç§¨ÃªÑ¡ µÑÇÍÂèÒ§
¤ÃÑ §à¡Ô´¤ÇÒÁ¤ÅÒ´à¤Å× Í¹ä»¨Ò¡µÓáË¹è§·Õ µéͧ¡ÒÃ
«Ö § ¨Ð·ÓãËé ¨Ñ§ËÇТͧàÇÅÒ·Õ ¨Ð·Ó¡Òêѡ µÑÇÍÂèÒ§áµèÅÐ
2.2.
ä·ÁÁÔ §ÃԤѿàÇÍÃÕà຺·Õ ãªé¡Ñ¹·Ñ Çä»
19
yk
received signal
y(t)
to data detection
t k = kT + τˆk timing tone detector
loop filter
TED
VCO
ÃÙ»·Õ 2.1: ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ¹ÔùÑÂ
ak
H(D)
rk
q(t)
τk
n(t) p(t)
y(t) LPF
yk = y (kT + τˆk )
t k = kT + τˆk
Viterbi detector
âk
z − d symbol detector
VCO D 1− D
loop filter
α+
β εˆk 1− D
rˆk −d
yk − d TED
ÃÙ»·Õ 2.2: Ẻ¨ÓÅͧªèͧÊÑÒ³ÍØ´Á¤µÔ ¾ÃéÍÁ¡Ñºä·ÁÁÔ §ÃԤѿàÇÍÃÕẺÍØ»¹ÑÂ
recovery) ÊÓËÃѺÃÒÂÅÐàÍÕ´à¡Õ ÂǡѺä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ¹ÔùÑÂÊÒÁÒöÈÖ¡ÉÒà¾Ô ÁàµÔÁä´é¨Ò¡ [20]
2.2
ä·ÁÁÔ §ÃԤѿàÇÍÃÕà຺·Õ ãªé¡Ñ¹·Ñ Çä»
¾Ô¨ÒóÒẺ¨ÓÅͧªèͧÊÑÒ³ÍØ´Á¤µÔ (ideal channel model) ã¹ÃÙ» ·Õ 2.2 ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ
ak ∈ {±1}
«Ö §ÁÕ¤ÒºàÇÅҢͧºÔµ
¤×Í ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô µÑÇ·Õ
k
T
¶Ù¡Ê觼èÒ¹ä»ÂѧªèͧÊÑÒ³
¢Í§ªèͧÊÑÒ³,
D
H(D) =
Pν
k=0 hk D
k â´Â·Õ
hk
¤×Í µÑÇ´Óà¹Ô¹¡ÒÃ˹èǧàÇÅÒ (delay operator), áÅÐ
20
ν
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¤×Í Ë¹èǤÇÒÁ¨Ó¢Í§ªèͧÊÑÒ³ ´Ñ§¹Ñ ¹ ÊÑÒ³ read back ÊÒÁÒöà¢Õ¹໠¹ÊÁ¡ÒÃä´é ¤×Í
p(t) =
X
rk q(t − kT − τk ) + n(t)
(2.1)
k â´Â·Õ
rk =
P
i ak−i hi ¤×Í ¢éÍÁÙÅ àÍÒµì¾Øµ ¢Í§ªèͧÊÑÒ³·Õ »ÃÒȨҡÊÑҳú¡Ç¹,
sin(πt/T )/(πt/T )
q(t) =
¤×Í ¿ §¡ìªÑ¹«Ô§¡ì (sinc function) ËÃ×ÍÊÑÒ³¾ÑÅÊì乤ÇÔµÊì (Nyquist pulse)
·Õ ÁÕ áº¹´ìÇÔ´·ì à¡Ô¹ à» ¹ ÈÙ¹Âì (zero excess bandwidth) [16], (unknown timing o set) µÑÇ ·Õ
k,
áÅÐ
n(t)
τk
¤×Í ÍͿ૵·Ò§àÇÅÒ·Õ äÁè ·ÃÒº¤èÒ
¤×Í ÊÑҳú¡Ç¹à¡ÒÊì ÊÕ ¢ÒÇẺºÇ¡ (AWGN:
additive white Gaussian noise) ·Õ ÁÕ ¤ÇÒÁ˹Òá¹è¹ Ê໡µÃÑÁ ¡ÓÅѧ (power spectrum density) ẺÊͧ´éÒ¹à·èÒ ¡Ñº 2
à´Ô¹áººÊØèÁ
N0 /2
ã¹Ë¹Ñ§Ê×Í ¹Õ ÍͿ૵·Ò§àÇÅÒ
τk
¨Ð¶Ù¡ ¨ÓÅͧãËé ÁÕ ÅѡɳÐà» ¹ ¡ÒÃ
(random walk) «Ö §¹ÔÂÒÁâ´Â [21]
τk+1 = τk + wk àÁ× Í
wk
¤×Í µÑÇá»ÃÊØèÁà¡ÒÊìà«Õ¹Ẻ
i.i.d. (independent and identically distributed) ·Õ ÁÕ¤èÒà©ÅÕ Â
(mean) à·èҡѺ¤èÒÈÙ¹Âì áÅÐÁÕ¤èÒ¤ÇÒÁá»Ã»Ãǹ (variance) à·èҡѺ
2) N (0, σw
â´Â¤èÒ
σw
(2.2)
2 σw
ËÃ×Íà¢Õ¹᷹ä´é´éÇÂ
¨Ðà» ¹µÑÇ¡Ó˹´ÃдѺ¤ÇÒÁÃعáç¢Í§ä·ÁÁÔ §¨ÔµàµÍÃì 3
·Õ ǧ¨ÃÀÒ¤ÃѺ ÊÑÒ³ read back ¨Ð¶Ù¡Ê觼èÒ¹ä»Âѧǧ¨Ã¡Ãͧ¼èÒ¹µ Ó ÍÔÁ¾ÑÅÊìà·èҡѺ
q(t)/T
wk ∼
¹Ñ ¹¤×Í ÁÕ¤ÇÒÁ¶Õ µÑ´ (cut o frequency) à·èҡѺ
(LPF) ·Õ Áռŵͺʹͧ
1/(2T )
à¾× Í·Ó˹éÒ·Õ ¡Ó¨Ñ´
ÊÑҳú¡Ç¹·Õ ÍÂÙè¹Í¡á¶º¤ÇÒÁ¶Õ (out of band noise) ¨Ò¡¹Ñ ¹ ¡ç¨Ð·Ó¡ÒêѡµÑÇÍÂèÒ§ÊÑÒ³ read back ·Õ àÇÅÒ
kT + τ̂k
à¾× Í·ÓãËéä´éà» ¹¢éÍÁÙÅá«Áà» Å
yk = y(kT + τ̂k ) =
X
ri q(kT + τ̂k − iT − τi ) + nk
(2.3)
i 2
Ẻ¨ÓÅͧ¡ÒÃà´Ô¹ ẺÊØèÁ ¹Õ ¶Ù¡ ¹Óãªé à¾ÃÒÐÇèÒ à» ¹ Ẻ¨ÓÅͧ·Õ §èÒ áÅÐÊÒÁÒöãªé á·¹ÅѡɳТͧªèͧÊÑÒ³
µèÒ§æ ä´é§èÒ â´Â¡ÒÃà»ÅÕ Â¹á»Å§¤èÒ¾ÒÃÒÁÔàµÍÃì 3
2 σw
à¾Õ§µÑÇà´ÕÂÇà·èÒ¹Ñ ¹
ÊÓËÃѺ Ãкº·Õ ÁÕ á¶º¤ÇÒÁ¶Õ ¨Ó¡Ñ´ (band limited system) ¹Ñ ¹¤×Í ¾Åѧ§Ò¹¢Í§ÊÑÒ³¨Ð¶Ù¡ ¨Ó¡Ñ´ ãËé ÍÂÙè 㹪èǧᶺ
¤ÇÒÁ¶Õ
|f | ≤ 1/(2T )
ǧ¨Ã¡Ãͧ¼èÒ¹µ Ó ¨Ð·ÓãËé ¢éÍÁÙÅ àÍÒµì¾Øµ ·Õ ä´é ÁÕ ¤èÒ ·Ò§Ê¶ÔµÔ ·Õ ¾Íà¾Õ§ (su cient statistic) [23]
àËÁ×͹¡Ñº¡ÒÃãªéǧ¨Ã¡ÃͧàËÁÒÐÊØ´ (matched lter) [16]
2.2.
ä·ÁÁÔ §ÃԤѿàÇÍÃÕà຺·Õ ãªé¡Ñ¹·Ñ Çä»
τ̂k
àÁ× Í
¤×Í ¤èÒ »ÃÐÁÒ³¢Í§
τk
ËÃ×Í ·Õ àÃÕ¡ÇèÒ ÍͿ૵·Ò§à¿Ê (phase o set) µÑÇ ·Õ
ªÑ¡ µÑÇÍÂèÒ§, áÅÐ
nk
N0 /(2T )
nk ∼ N (0, σn2 )
¹Ñ ¹¤×Í
21
k
¢Í§¡ÒÃ
¤×Í AWGN ·Õ ÁÕ ¤èÒà©ÅÕ Â à·èÒ ¡Ñº ¤èÒ ÈÙ¹Âì áÅФèÒ ¤ÇÒÁá»Ã»Ãǹà·èÒ ¡Ñº
σn2 =
ǧ¨Ã TED ¨Ð·Ó˹éÒ·Õ ã¹¡ÒäӹdzËÒ¤èÒ»ÃÐÁÒ³¢Í§¢éͼԴ¾ÅÒ´·Ò§àÇÅÒ (timing error) =
τk − τ̂k
²k
«Ö §¡ç¤×Í ¤èÒ¤ÇÒÁäÁèµÃ§¡Ñ¹ÃÐËÇèÒ§à¿Ê¢Í§ÊÑÒ³á͹ÐÅçÍ¡·Õ ä´éÃѺ¡Ñºà¿Ê¢Í§ÊÑÒ³
¹ÒÌ ¡Ò¢Í§Ç§¨Ã PLL ã¹·Ò§»¯ÔºÑµÔ áÅéÇ Ç§¨Ã TED ÁÕ ËÅÒ»ÃÐàÀ· [4] â´Â¨Ð¢Ö ¹ ÍÂÙè ¡Ñº ÅѡɳР¡ÒùӢéÍÁÙÅ·Õ ´éÒ¹¢Òà¢éҢͧǧ¨Ã TED ÁÒãªé§Ò¹ «Ö §â´Â·Ñ Çä»áÅéÇ »ÃÐÊÔ·¸ÔÀÒ¾¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕ ¨Ð¢Ö ¹ ÍÂÙè ¡Ñº ¤Ø³ÀÒ¾¢Í§Ç§¨Ã TED ã¹Ë¹Ñ§Ê×Í àÅèÁ ¹Õ ¨Ð¾Ô¨ÒóÒ੾ÒÐǧ¨Ã TED ·Õ ¹ÔÂÁãªé §Ò¹¡Ñ¹ ¹Ñ ¹¤×Í Ç§¨Ã TED Ẻ Mueller and Müller ËÃ×ÍàÃÕ¡ÊÑ ¹æ ÇèÒ M&M TED [24] «Ö §¨Ð¤Ó¹Ç³ ËÒ¤èÒ»ÃÐÁÒ³¢Í§¢éͼԴ¾ÅÒ´·Ò§àÇÅÒ ´Ñ§¹Õ
²̂k = KT {yk r̂k−1 − yk−1 r̂k } â´Â·Õ =
²
r̂k
¤×Í ¤èÒ»ÃÐÁÒ³¢Í§
rk , KT
(2.4)
¤×Í ¤èÒ¤§µÑÇ (constant) ·Õ ¶Ù¡ãªéà¾× Í·ÓãËéÁÑ ¹ã¨ä´éÇèÒ
E[²̂k |²]
àÁ× Í ÃдѺ ÍѵÃÒÊèǹ¤èÒ ¡ÓÅѧ à©ÅÕ Â ¢Í§ÊÑÒ³·Õ µéͧ¡ÒõèÍ ¤èÒ ¡ÓÅѧ à©ÅÕ Â ¢Í§ÊÑҳú¡Ç¹
(SNR: signal to noise ratio) ÁÕ¤èÒÊÙ§ ËÃ×ÍÍÒ¨¨Ð¡ÅèÒÇä´éÇèÒ ¤èÒ
KT
¶Ù¡¹ÓÁÒãªéà¾× Í·ÓãËé¤ÇÒÁªÑ¹
¢Í§àÊé¹â¤é§ÃÙ»µÑÇàÍÊ (S curve) [4] ÁÕ¤èÒà·èҡѺ¤èÒË¹Ö § ³ ¨Ø´¡Óà¹Ô´, áÅÐ
E[·]
¤×Í µÑÇ´Óà¹Ô¹¡ÒÃ
¤èÒ ¤Ò´ËÁÒ (expectation operator) [10, 25, 26] ¨Ò¡ÊÁ¡Òà (2.4) ¨ÐàËç¹ä´éÇèÒ »ÃÐÊÔ·¸ÔÀÒ¾ ¢Í§Ç§¨Ã TED ¨Ð¢Ö ¹ ÍÂÙè ¡Ñº ¤èÒ µÑ´ÊԹ㨠(decision)
{r̂k }
´Ñ§¹Ñ ¹ »ÃÐÊÔ·¸ÔÀÒ¾¢Í§ä·ÁÁÔ § ÃԤѿàÇÍÃÕ
¨Ðà» ¹¿ §¡ìªÑ¹¢Í§¤ÇÒÁ¹èÒàª× Ͷ×ͧ͢¤èҵѴÊÔ¹ã¨áÅÐ SNR ·Õ ãªé ¨Ö§à» ¹à˵ؼÅÇèÒ ·ÓäÁǧ¨ÃµÃǨËÒ 4
ÊÑÅѡɳì (symbol detector) ·Õ ãªéã¹ä·ÁÁÔ §ÃԤѿàÇÍÃÕ ¤×Í Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ [15] ·Õ ÁÕ»ÃÔÁÒ³¡ÒÃ˹èǧàÇÅÒÊÓËÃѺ¡ÒõѴÊԹ㨠(decision delay) à·èҡѺ
dT
(Viterbi detector)
˹èÇ (àªè¹
d = 4)
á·¹¡ÒÃãªé§Ò¹Ç§¨ÃµÃǨËÒ¢Õ´àÃÔ Áà»ÅÕ Â¹áººËÅÒÂÃдѺ (multi level threshold detector) [27] ËÅѧ¨Ò¡¹Ñ ¹ ¤èÒ»ÃÐÁÒ³¢Í§¢éͼԴ¾ÅÒ´·Ò§àÇÅÒ
²̂k
¨Ð¶Ù¡Ê觼èÒ¹ä»Âѧǧ¨Ã¡ÃͧÅÙ» à¾× ͡ӨѴ
ÊÑҳú¡Ç¹·Õ ὧÍÂÙè ã¹ÊÑÒ³¢éͼԴ¾ÅÒ´·Ò§àÇÅÒ áÅÐÍͿ૵·Ò§à¿Ê¢Í§¡Òêѡ µÑÇÍÂèÒ§ (sampling phase o set) µÑÇ ¶Ñ´ ä» 4
τ̂k+1
¡ç ¨Ð¶Ù¡ »ÃѺ ¤èÒ (update) â´Âǧ¨Ã PLL Íѹ´Ñº ·Õ Êͧ
ÈÖ¡ÉÒÃÒÂÅÐàÍÕ´ËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ä´é㹺··Õ 4
22
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
(second order PLL) µÒÁ¤ÇÒÁÊÑÁ¾Ñ¹¸ì´Ñ§¹Õ [4]
àÁ× Í
θ̂k
θ̂k+1 = θ̂k + β²̂k ,
(2.5)
τ̂k+1 = τ̂k + α²̂k + θ̂k+1
(2.6)
¤×Í ¤èÒ »ÃÐÁÒ³¢Í§¢éͼԴ¾ÅÒ´·Ò§¤ÇÒÁ¶Õ (frequency error) [27], áÅÐ
α
áÅÐ
β
¤×Í
¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL [4] «Ö §¨Ðà» ¹µÑÇ¡Ó˹´ ẹ´ìÇÔ´·ì¢Í§ÅÙ» (loop bandwidth) áÅÐÍѵÃÒ ¡ÒÃÅÙèà¢éÒ (convergence rate) ¡ÅèÒǤ×Í ¶éÒ¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL ÂÔ §ÁÒ¡ ẹ´ìÇÔ´·ì¢Í§ÅÙ»¡ç ¨Ð¡ÇéÒ§ «Ö §¨Ð·ÓãËéÍѵÃÒ¡ÒÃÅÙèà¢éÒ¡ç¨ÐàÃçÇ áµèÊÑҳú¡Ç¹·Õ à¢éÒÁÒã¹Ç§¨Ã PLL ¡ç¨ÐÁÕÁÒ¡ ÊÓËÃѺ ã¹¡Ã³Õ·Õ ÃкºÁÕ੾ÒТéͼԴ¾ÅÒ´·Ò§à¿Ê (phase error) à·èÒ¹Ñ ¹ ǧ¨ÃÀÒ¤ÃѺÍÒ¨¨Ð¹Óǧ¨Ã PLL Íѹ´Ñº·Õ Ë¹Ö § ( rst order PLL) ÁÒãªé§Ò¹á·¹Ç§¨Ã PLL Íѹ´Ñº·Õ Êͧ¡çä´é â´Â·Õ ÍͿ૵·Ò§à¿Ê ¢Í§¡ÒêѡµÑÇÍÂèÒ§µÑǶѴä»
τ̂k+1
¨Ð¶Ù¡»ÃѺ¤èÒµÒÁ¤ÇÒÁÊÑÁ¾Ñ¹¸ì´Ñ§¹Õ [4]
τ̂k+1 = τ̂k + α²̂k
(2.7)
â´Â·Ñ Çä» ä·ÁÁÔ §ÃԤѿàÇÍÃըзӧҹ໠¹ 2 ÀÒÇÐ (mode) ¤×Í
1) ÀÒÇСÒÃä´é ÁÒ (acquisition mode) ¨Ð·Ó§Ò¹ã¹µÍ¹àÃÔ Áµé¹ ¢Í§¡Ãкǹ¡ÒÃà¢éÒ ¨Ñ§ËÇдéÇ ¤ÇÒÁªèÇÂàËÅ×Í ¢Í§áºº¢éÍÁÙÅ (data pattern) ·Õ àÃÕ¡ÇèÒ preamble [27] à¹× ͧ¨Ò¡ ä·ÁÁÔ § ÃԤѿàÇÍÃÕ ÃÙé á¹è¹Í¹ÇèÒ preamble ÁÕ ÅѡɳÐà» ¹ ÍÂèÒ§äà ¨Ö§ ·ÓãËé ÊÒÁÒö·ÃÒºä´é ÇèÒ ¤èÒ ¶Ù¡µéͧ¤×ͤèÒÍÐäà ´Ñ§¹Ñ ¹ã¹ªèǧÀÒÇСÒÃä´éÁÒ¹Õ Ç§¨Ã PLL ¨Ðãªé¤èÒ ËÒ¤èÒ
²̂k
r̂k = rk
r̂k
·Õ
㹡Òäӹdz
µÒÁÊÁ¡Òà (2.4) (ǧ¨ÃµÃǨËÒÊÑÅѡɳì·Õ ãªéã¹ä·ÁÁÔ §ÃԤѿàÇÍÃÕ¨ÐÂѧäÁè¶Ù¡ãªé§Ò¹
㹪èǧ¹Õ ) «Ö § ¨Ð·ÓãËé ä´é ¤èÒ ·Õ ¶Ù¡µéͧ à¾ÃÒÐ©Ð¹Ñ ¹ ¡Ãкǹ¡ÒÃä·ÁÁÔ § ÃԤѿàÇÍÃÕ ã¹ªèǧ¹Õ ¨Ö§ ÁÕ ¤ÇÒÁ¹èÒàª× Ͷ×ÍÁÒ¡ â´Â¨Ø´»ÃÐʧ¤ìËÅÑ¡¢Í§ÀÒÇСÒÃä´éÁÒ ¡ç¤×Í ¡ÒÃËÒ¤èÒ»ÃÐÁÒ³àÃÔ Áµé¹¢Í§ ÍͿ૵·Ò§à¿ÊáÅÐÍͿ૵·Ò§¤ÇÒÁ¶Õ (frequency o set) ·Õ ὧÍÂÙè ã¹ÊÑÒ³á͹ÐÅçÍ¡ ·Õ ¨Ð·Ó¡ÒêѡµÑÇÍÂèÒ§
2) ÀÒÇСÒõԴµÒÁ (tracking mode) ¨Ð·Ó§Ò¹µèͨҡÀÒÇСÒÃä´éÁÒ â´Âã¹¢Ñ ¹µÍ¹¹Õ ¤èÒ ãªé㹡ÒäӹdzËÒ¤èÒ
²̂k
r̂k
·Õ
µÒÁÊÁ¡Òà (2.4) ¨Ðä´éÁҨҡǧ¨ÃµÃǨËÒÊÑÅѡɳì·Õ ãªéã¹ä·ÁÁÔ §
ÃԤѿàÇÍÃÕ («Ö §ÍÒ¨¨ÐÁդسÀÒ¾äÁè´Õ àÁ× Íà·Õº¡Ñº¡ÒÃãªé
rk
¨ÃÔ§æ) ´Ñ§¹Ñ ¹ ¨Ø´»ÃÐʧ¤ìËÅÑ¡¢Í§
2.3.
¡ÒÃÍÍ¡à຺¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL
23
ÀÒÇСÒõԴµÒÁ ¡ç ¤×Í à» ¹ ¡ÒÃá¡éä¢áÅлÃѺ»Ãا ¤èÒ àÃÔ Áµé¹ ¢Í§ÍͿ૵·Ò§à¿ÊáÅÐÍͿ૵ ·Ò§¤ÇÒÁ¶Õ ·Õ ä´éÁÒ¨Ò¡ÀÒÇСÒÃä´éÁÒ
¨ÐàËç¹ä´éÇèÒ㹪èǧÀÒÇСÒÃä´éÁÒ Ç§¨Ã PLL ·ÃÒºá¹è¹Í¹ÇèÒ preamble ¤×ÍÍÐäà ´Ñ§¹Ñ ¹ ǧ¨Ã PLL ¨Ö§ÊÒÁÒö·Õ ¨Ðãªé¤èÒ
α
áÅÐ
β
α
áÅÐ
β
·Õ ÁÕ¤èÒÁÒ¡ä´é à¾× ͪèÇ·ÓãËéÁÕÍѵÃÒ¡ÒÃÅÙèà¢éÒ·Õ ÃÇ´àÃçÇ ÍÂèÒ§äáçµÒÁ ¤èÒ
·Õ ãªé¤ÇÃ·Õ ¨ÐÁÕ¤èÒŴŧ àÁ× Íä·ÁÁÔ §ÃԤѿàÇÍÃÕà¢éÒÊÙèªèǧÀÒÇСÒõԴµÒÁ à¾× ÍÅ´¼Å¡Ãзº¢Í§
ÊÑҳú¡Ç¹·Õ ¨Ðà¢éÒÁÒã¹Ç§¨Ã PLL [28] ´Ñ§¹Ñ ¹ ¹Ñ¡Í͡ẺÃкº¨Ðµéͧ»ÃйջÃйÍÁÃÐËÇèÒ§ ẹ´ìÇÔ´·ì¢Í§ÅÙ»áÅÐÍѵÃÒ¡ÒÃÅÙèà¢éÒ ã¹ÃÐËÇèÒ§·Õ ·Ó¡ÒÃÍ͡Ẻ¤èÒ¾ÒÃÒÁÔàµÍÃì
2.3
α
áÅÐβ
¡ÒÃÍÍ¡à຺¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL
¨Ò¡¼Å¡Ò÷´Åͧ·Õ ä´é ÃѺ ¨Ò¡¡ÒèÓÅͧÃкº (system simulation) ¾ºÇèÒ ÇÔ¸Õ¡ÒÃ·Õ ´Õ ·Õ ÊØ´ 㹡Òà Í͡Ẻ¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL (·Ñ §
α áÅÐ β ) ¤×Í ¡ÒÃàÅ×Í¡ãªé¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL ·Õ
·ÓãËéÃкºÁÕ BER ¹éÍÂÊØ´ ³ ·Õ ǧ¨ÃÀÒ¤ÃѺ ÍÂèÒ§äáçµÒÁ ÇÔ¸Õ¹Õ äÁèÊÒÁÒöãªé§Ò¹ä´é¨Ãԧ㹷ҧ»¯ÔºÑµÔ à¹× ͧ¨Ò¡ µéͧãªé ÃÐÂÐàÇÅҹҹ㹡ÒÃ·Õ ¨ÐËÒ¤èÒ ¾ÒÃÒÁÔàµÍÃì ¢Í§Ç§¨Ã PLL ·Õ àËÁÒÐÊØ´ ´Ñ§¹Ñ ¹ â´Â ·Ñ Çä»áÅéÇ áºº¨ÓÅͧ¡Ò÷ӧҹ¢Í§Ç§¨Ãà¿ÊÅçÍ¡ÅٻẺàªÔ§àÊé¹ÁÑ¡¨Ð¶Ù¡¹ÓÁÒãªé㹡ÒÃÍ͡Ẻ¤èÒ ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL [4] ࡳ±ì (criterion) Ë¹Ö §·Õ à» ¹ä»ä´é ¡ç¤×Í ¡ÒÃàÅ×Í¡ãªé¤èÒ
α
áÅÐ
β
·Õ
·ÓãËé ¼ÅµÍºÊ¹Í§¢Í§Ãкº (system response) ÊÒÁÒö·Õ ¨ÐµÒÁ·Ñ¹ ¡ÒÃà»ÅÕ Â¹á»Å§¢Í§à¿ÊáÅÐ ¤ÇÒÁ¶Õ ¢Í§ÊÑÒ³ read back ÀÒÂ㹪èǧ C ºÔµ ËÃ×Í ¤ÒºàÇÅҢͧºÔµ (bit period) «Ö § ÍÒ¨¨Ð ¾Ô¨ÒóÒä´éÇèÒ à¡³±ì¹Õ ÊÍ´¤Åéͧ¡ÑºÍѵÃÒ¡ÒÃÅÙèà¢éÒ ¡ÅèÒǤ×Í ¶éÒ¤èÒ
C
ÂÔ §¹éÍ ¡çËÁÒ¤ÇÒÁÇèÒ ä·ÁÁÔ §
ÃԤѿàÇÍÃÕ¨ÐÁÕÍѵÃÒ¡ÒÃÅÙèà¢éÒ·Õ ÂÔ §àÃçÇ
2.3.1
¡ÒÃÇÔà¤ÃÒÐËìàªÔ§àÊ鹢ͧǧ¨Ã PLL Íѹ´Ñº·Õ Ë¹Ö §
ǧ¨Ã PLL Íѹ´Ñº·Õ Ë¹Ö § ( rst order PLL) ¨ÐÊÒÁÒö¨Ñ´¡ÒáѺ¢éͼԴ¾ÅÒ´·Ò§à¿Êä´éà¾Õ§ÍÂèÒ§à´ÕÂÇ (äÁè ÊÒÁÒö¨Ñ´¡ÒáѺ ¢éͼԴ¾ÅÒ´·Ò§¤ÇÒÁ¶Õ ä´é) ã¹Êèǹ¹Õ ¨Ð͸ԺÒ¡ÒÃÍ͡Ẻ¤èÒ ¾ÒÃÒÁÔàµÍÃì ¢Í§ ǧ¨Ã PLL Íѹ´Ñº·Õ Ë¹Ö § «Ö §¨Ðà» ¹»ÃÐ⪹ìµèÍ¡ÒÃàÃÕ¹ÃÙéǧ¨Ã PLL Íѹ´ÑºÊÙ§æ µèÍä» ÊÓËÃѺ㹡Òà ÇÔà¤ÃÒÐËì¹Õ ¢éͼԴ¾ÅÒ´·Ò§à¿Ê¨Ð¶Ù¡¨ÓÅͧãËéà» ¹¿ §¡ìªÑ¹¢Ñ ¹ºÑ¹ä´ (step function) ¹Ñ ¹¤×Í
τk = T
24
àÁ× Í
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
k≥0
áÅÐ
τk = 0
àÁ× Í
k<0
¾Ô¨ÒóÒÊÁ¡ÒûÃѺ¤èÒÍͿ૵·Ò§à¿Ê¢Í§¡ÒêѡµÑÇÍÂèÒ§µÑǶѴä»
τ̂k+1
¢Í§Ç§¨Ã PLL Íѹ´Ñº
·Õ Ë¹Ö § ´Ñ§µèÍ仹Õ
τ̂k+1 = τ̂k + α²̂k−d àÁ× Í
d
(2.8)
¤×Í »ÃÔÁҳ˹èǧàÇÅÒã¹ä·ÁÁÔ §ÅÙ» (timing loop) ÁÕ˹èÇÂà» ¹ºÔµà«ÅÅì
¤èÒ»ÃÐÁÒ³¢Í§¢éͼԴ¾ÅÒ´·Ò§àÇÅÒ, timing error), áÅÐ
vk
²k
=
τk − τ̂k
T , ²̂k
=
²k + vk
¤×Í
¤×Í ¢éͼԴ¾ÅÒ´·Ò§àÇÅÒ·Õ ËŧàËÅ×ÍÍÂÙè (residual
¤×Í ÊÑҳú¡Ç¹ã¹Ç§¨Ã TED ¶éÒÊÁÁصÔãËé
vk
ÁÕ¤èÒ¹éÍÂÁÒ¡ (ÊÒÁÒö·Õ
¨Ðà¾Ô¡à©Âä´é) ´Ñ§¹Ñ ¹ ¿ §¡ìªÑ¹¶èÒÂâ͹ (transfer function) [12] ¢Í§ÃкºµÒÁÊÁ¡Òà (2.8) ÊÒÁÒö ·Õ ¨ÐËÒä´éâ´Â¡ÒÃãªé¡ÒÃá»Å§«Õ (Z transform) [12, 16] ¹Ñ ¹¤×Í
G(z) = àÁ× Í
Γ̂(z)
áÅÐ
¢éͼԴ¾ÅÒ´
²k
Γ(z) =
Γ̂(z) αz −(d+1) = Γ(z) 1 − z −1 + αz −(d+1)
¤×Í ¼Å¡ÒÃá»Å§«Õ¢Í§
τk − τ̂k
τ̂k
áÅÐ
E(z)
µÒÁÅӴѺ à¾ÃÒÐ©Ð¹Ñ ¹ ¿ §¡ìªÑ¹¶èÒÂâ͹¢Í§
ÊÒÁÒöà¢Õ¹ä´éà» ¹
E(z) = Γ(z) − Γ̂(z) = àÁ× Í
τk
(2.9)
¤×Í ¼Å¡ÒÃá»Å§«Õ¢Í§
1 − z −1 Γ(z) 1 − z −1 + αz −(d+1)
(2.10)
²k
ࡳ±ì ·Õ ÍÒ¨¨Ð¹ÓÁÒãªé 㹡ÒÃàÅ×Í¡¤èÒ
α
¤×Í ¡ÒÃàÅ×Í¡¤èÒ
α
·Õ ·ÓãËé àʶÕÂÃÀÒ¾ (stable) ¢Í§
ÃкºáÅÐÍѵÃÒ¡ÒÃÅÙèà¢éÒà» ¹·Õ ¹èÒ¾Í㨠ÊÓËÃѺáµèÅФèҢͧ»ÃÔÁҳ˹èǧàÇÅÒã¹ÅÙ»
d
(loop delay)
·Õ ¡Ó˹´ãËéÁÒ ÇÔ¸Õ¡ÒÃ¹Õ ÊÒÁÒö·Óä´éâ´Âãªé·Ñ §ÊÁ¡Òà (2.9) ËÃ×Í (2.10) ¡çä´é ¡ÅèÒǤ×Í ¢Ñ ¹µÍ¹áá ãËé ËÒ¤èÒ
α
·Ñ §ËÁ´·Õ ·ÓãËé ÃкºÁÕ ¤ÇÒÁàʶÕÂÃÀÒ¾ ¨Ò¡ÊÁ¡Òà (2.9) áÅÐ (2.10) Ãкº¨ÐÁÕ ¤ÇÒÁ
àʶÕÂÃÀÒ¾ ¡ç µèÍàÁ× Í ·Ø¡ â¾Å (all poles) ËÃ×Í ÃÒ¡¤ÓµÍº (root) ¢Í§µÑÇ Êèǹ ¢Í§ÊÁ¡Òà (2.9) ËÃ×Í (2.10) ÍÂÙèÀÒÂã¹Ç§¡ÅÁË¹Ö §Ë¹èÇ [16] ¨Ò¡¡ÒÃá¡éÊÁ¡ÒèÐä´éÇèÒ ¤èÒ
α
·Õ ·ÓãËéÃкºÁÕ¤ÇÒÁ
àʶÕÂÃÀÒ¾ ÊÒÁÒöËÒä´é¨Ò¡ [4]
µ 0 < α < 2 sin ÃÙ» ·Õ 2.3 áÊ´§¤èÒ ÁÒ¡ÊØ´ ¢Í§
α
π 4d + 2
¶
·Õ Âѧ¤§·ÓãËé ÃкºÁÕ ¤ÇÒÁàʶÕÂÃÀÒ¾ ÊÓËÃѺ áµèÅФèÒ
(2.11)
d
¨ÐàËç¹
2.3.
¡ÒÃÍÍ¡à຺¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL
25
2
α (PLL gain parameter)
1.8 1.6 1.4 1.2 1 0.8 0.6 0.4
α
max
0.2 0
0
10
20
30
40
50
Normalized loop delay, d
ÃÙ»·Õ 2.3: ¤èÒÁÒ¡ÊØ´¢Í§
α
ä´é ªÑ´à¨¹ÇèÒ ªèǧàʶÕÂÃÀÒ¾¢Í§¤èÒ ¤èÒ
α
·Õ Âѧ¤§·ÓãËéÃкºÁÕ¤ÇÒÁàʶÕÂÃÀÒ¾ ÊÓËÃѺáµèÅФèÒ
α
¨ÐŴŧÍÂèÒ§ÃÇ´àÃçÇ àÁ× Í
ËÅÒ¤èÒ ·Õ ·ÓãËé ÃкºÁÕ ¤ÇÒÁàʶÕÂÃÀÒ¾ áµè ¨ÐàÅ×Í¡ãªé
α
d
ÁÕ ¤èÒ à¾Ô Á ¢Ö ¹ áÅж֧áÁéÇèÒ ¨ÐÁÕ
à¾Õ§¤èÒ à´ÕÂÇ·Õ ·ÓãËé ¼ÅµÍºÊ¹Í§
¢Í§ÃкºÊÒÁÒö·Õ ¨ÐµÔ´µÒÁ¼ÅµÍºÊ¹Í§¢Ñ ¹ºÑ¹ä´ (step response) ÀÒÂã¹ ¤ÇÒÁ¤ÅÒ´à¤Å× Í¹ä´é
α
±5%
C
ºÔµ â´ÂÂÍÁãËé ÁÕ
â´Â·Õ µÑÇàÅ¢ 5% ¹Õ ¶Ù¡ ¹Óãªé à¾× Í à» ¹ ¡Òüè͹»Ã¹à¡³±ì ¡ÒÃÍ͡Ẻ
à¾× Í·Õ ¨Ðä´éÅ´¼Å¡Ãзº¢Í§ÊÑҳú¡Ç¹·Õ ¨Ðà¢éÒÁÒã¹Ç§¨Ã PLL ÃÙ»·Õ 2.4(a) áÊ´§¤èÒ
·ÓãËéÃкºÁÕ¤ÇÒÁàʶÕÂÃÀÒ¾áÅÐÁÕÍѵÃÒ¡ÒÃÅÙèà¢éÒÀÒÂã¹ (¹Ñ ¹¤×Í ¤èÒ
C
¾ºÇèÒÁÕà¾Õ§ ÊÓËÃѺ
d
α
ÁÕ¤èÒ¹éÍÂ) ¤èÒ
α
C
¡ç¨ÐÂÔ §ÁÕ¤èÒÁÒ¡ ¨Ò¡·Õ áÊ´§ã¹ÃÙ»·Õ 2.4(a) àÁ× Í¡Ó˹´
ºÒ§¤èÒ·Õ ·ÓãËéÃкºÁÕ¤ÇÒÁàʶÕÂÃÀÒ¾ÊÓËÃѺ¤èÒ
µÑ §áµè ¤èÒ 0 ¶Ö§
30T
= 0 ¶Ö§
30T
à¹× ͧ¨Ò¡
·Õ
d
C
ÁÒãËé ¨Ð
¤èÒË¹Ö § µÑÇÍÂèÒ§àªè¹ ¨ÐÁÕ¤èÒ
α100
à·èÒ¹Ñ ¹ ·Õ ·ÓãËé ÃкºÁÕ ¤ÇÒÁàʶÕÂÃÀÒ¾áÅÐÁÕ ÍѵÃÒ¡ÒÃÅÙèà¢éÒ ÀÒÂã¹
«Ö §ÊÍ´¤Åéͧ¡Ñºà¡³±ì¡ÒÃÍ͡Ẻ
α100
αC
ºÔµ Êѧࡵ¨Ð¾ºÇèÒ ÍѵÃÒ¡ÒÃÅÙèà¢éÒÂÔ §àÃçÇ
100 ºÔµ ÃÙ»·Õ 2.4(b) áÊ´§¼ÅµÍºÊ¹Í§¢Ñ ¹ºÑ¹ä´¢Í§ÃкºµÒÁÊÁ¡Òà (2.9) â´Âãªé
d
d
α
α100
ÊÓËÃѺ
·Õ µÑ §äÇé
¶Ù¡Í͡ẺÁÒà¾× ÍãËéÃкºÊÒÁÒö·Õ ¨ÐµÔ´µÒÁ¼ÅµÍºÊ¹Í§¢Ñ ¹ºÑ¹ä´ÀÒÂã¹ 100
ºÔµ ´éǤèÒ¤ÇÒÁ¤ÅÒ´à¤Å× Í¹ÂÔ¹ÂÍÁ (tolerance)
±5% ´Ñ§¹Ñ ¹ ¤èÒÊÑÁºÙóì¢Í§¢¹Ò´¢Í§¼ÅµÍºÊ¹Í§
26
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
0.05
α (PLL gain parameter)
0.045 0.04 0.035 0.03
α
0.025
64
0.02
0.01 0.005
α
α200
0.015
100
α300 0
10
20
30
40
50
(a) Normalized loop delay, d
d increases 1.05 0.95
Magnitude
± 5% tolerance
d increases
Convergence rate within 100 samples
0
100
200
300
400
500
(b) Time (in bit periods)
ÃÙ»·Õ 2.4: (a) ¤èÒ
d,
αC
·Õ ·ÓãËéÃкºÁÕ¤ÇÒÁàʶÕÂÃÀÒ¾áÅÐÁÕÍѵÃÒ¡ÒÃÅÙèà¢éÒÀÒÂã¹
áÅÐ (b) ¼ÅµÍºÊ¹Í§¢Í§ÃкºàÁ× Íãªé
¢éͼԴ¾ÅÒ´ (error response)
E(z)
α100
ÊÓËÃѺ¤èÒ
d
¨Ò¡ 0 ¶Ö§
C
ºÔµ ÊÓËÃѺáµèÅÐ
30T
µÒÁÊÁ¡Òà (2.10) ¤ÇÃ·Õ ¨ÐÁÕ ¢¹Ò´¹éÍ¡ÇèÒ 0.05 ËÅѧ¨Ò¡·Õ
¢éÍÁÙżèÒ¹ä» 100 ºÔµ µÒÁ·Õ áÊ´§ã¹ÃÙ»·Õ 2.5 ÊÓËÃѺ
d = 14T
Êѧࡵ¨Ð¾ºÇèÒ ÁÕ¤èÒ
α100
¨Ó¹Ç¹
2.3.
¡ÒÃÍÍ¡à຺¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL
1.4
27
α = 0.055 α = 0.035
Magnitude
1.05 1 0.95
α = 0.0218 α = 0.01 α = 0.005
0
0
14
50
100
150
200
250
(a) Time (in bit periods) 1
Magnitude
α = 0.005 α = 0.0218
α = 0.01
0.05 0 −0.05
α = 0.035 α = 0.055 −0.4
0
50
100
150
200
250
(b) Time (in bit periods)
ÃÙ»·Õ 2.5: (a) ¼ÅµÍºÊ¹Í§¢Ñ ¹ºÑ¹ä´¢Í§Ãкº áÅÐ (b) ¼ÅµÍºÊ¹Í§¢éͼԴ¾ÅÒ´ÊÓËÃѺ áÅФèÒ
α
2 ¤èÒ ¤×Í
d = 14T
µèÒ§æ
α = 0.0218
¤ÅÒ´à¤Å× Í¹ÂÔ¹ÂÍÁ
áÅÐ
±5%
α = 0.055
·Õ ·ÓãËéÃкºÁÕÍѵÃÒ¡ÒÃÅÙèà¢éÒÀÒÂã¹ 100 ºÔµ ´éǤèÒ¤ÇÒÁ
ÍÂèÒ§äáçµÒÁã¹·Ò§»¯ÔºÑµÔ ¤èÒ
α
·Õ ÁÕ¤èÒ¹éÍ¡ÇèÒ ¨Ð¶Ù¡àÅ×Í¡ÁÒãªé§Ò¹ã¹
28
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ǧ¨Ã PLL à¾× Í·Õ ¨Ð·ÓãËéẹ´ìÇÔ´·ì¢Í§ÅÙ»ÁÕ¤èÒ¹éÍ «Ö §¨ÐªèÇÂÅ´¼Å¡Ãзº¢Í§ÊÑҳú¡Ç¹·Õ ¨Ð à¢éèÒÁÒã¹Ç§¨Ã PLL ä´é
2.3.2
¡ÒÃÇÔà¤ÃÒÐËìàªÔ§àÊ鹢ͧǧ¨Ã PLL Íѹ´Ñº·Õ Êͧ
ã¹¡Ã³Õ ·Õ ÃкºÁÕ Í§¤ì»ÃСͺ¢Í§ÍͿ૵·Ò§¤ÇÒÁ¶Õ (frequency o set) ǧ¨Ã PLL Íѹ´Ñº ·Õ Êͧ áÅÐ
β
ÊÓËÃѺ¤èÒ
d
¨Ðµéͧ¶Ù¡¹ÓÁÒãªé§Ò¹á·¹Ç§¨Ã PLL Íѹ´Ñº·Õ Ë¹Ö § «Ö §¾ÒÃÒÁÔàµÍÃì·Õ ¨Ðµéͧ¤Ó¹Ç³ËÒ ¤×Í ã¹¡ÒÃÍ͡Ẻ¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL Íѹ´Ñº·Õ Êͧ¹Õ ¨ÐàÃÔ Áµé¹¨Ò¡¡ÒÃËÒ¤èÒ áÅÐ
C
α
α
·Õ ¡Ó˹´ÁÒãËé â´ÂÊÁÁصÔÇèÒ ã¹ÃкºÁÕà¾Õ§á¤è¢éͼԴ¾ÅÒ´·Ò§à¿Êà·èÒ¹Ñ ¹ (¹Ñ ¹¤×Í ãªéÇÔ¸Õ¡ÒÃ
Í͡Ẻ¤èÒ
α
µÒÁ·Õ ͸ԺÒÂã¹ËÑÇ¢éÍ ·Õ 2.3.1) ¨Ò¡¹Ñ ¹ àÁ× Í ä´é ¤èÒ
α
·Õ µéͧ¡ÒÃáÅéÇ ¡ç ¨ÐàÃÔ Á ËÒ¤èÒ
β
â´Âãªé¡ÒÃÇÔà¤ÃÒÐËìàªÔ§àÊ鹢ͧǧ¨Ã PLL Íѹ´Ñº·Õ Êͧ ÊÓËÃѺ»ÃÔÁÒ³¢Í§ÍͿ૵·Ò§¤ÇÒÁ¶Õ ·Õ ãËéÁÒ
τk
=
¢Í§ÍͿ૵·Ò§¤ÇÒÁ¶Õ ÁÕ˹èÇÂà» ¹à»ÍÃìà«ç¹µì (percent) ¢Í§ºÔµà«ÅÅì
T
«Ö §ã¹¡ÒÃÇÔà¤ÃÒÐËì¹Õ ¢éͼԴ¾ÅÒ´·Ò§¤ÇÒÁ¶Õ ¨Ð¶Ù¡¨ÓÅͧãËéÁÕ¤èÒà» ¹
¾Ô¨ÒóÒÊÁ¡ÒûÃѺ¤èÒÍͿ૵·Ò§à¿Ê¢Í§¡ÒêѡµÑÇÍÂèÒ§µÑǶѴä»
kfd
τ̂k+1
àÁ× Í
fd
¤×Í »ÃÔÁÒ³
¢Í§Ç§¨Ã PLL Íѹ´Ñº
·Õ Êͧ ´Ñ§µèÍ仹Õ
θ̂k+1 = θ̂k + β²̂k−d
(2.12)
τ̂k+1 = τ̂k + α²̂k−d + θ̂k+1
(2.13)
àªè¹à´ÕÂǡѹ ¶éÒÊÁÁصÔÇèÒäÁèÁÕÊÑҳú¡Ç¹ã¹Ç§¨Ã TED ¹Ñ ¹¤×Í
²̂k
=
²k
=
τk − τ̂k ,
à¾ÃÒÐ©Ð¹Ñ ¹
¿ §¡ìªÑ¹ ¶èÒÂâ͹¢Í§ÃкºµÒÁÊÁ¡Òà (2.12) áÅÐ (2.13) ÊÒÁÒö·Õ ¨ÐËÒä´é â´Â¡ÒÃãªé ¡ÒÃá»Å§«Õ ´Ñ§¹Õ
G(z) =
Γ̂(z) (α + β)z −(d+1) − αz −(d+2) = Γ(z) 1 − 2z −1 + z −2 + (α + β)z −(d+1) − αz −(d+2)
(2.14)
áÅп §¡ìªÑ¹¶èÒÂâ͹¢Í§¢éͼԴ¾ÅÒ´ ¤×Í
E(z) = Γ(z) − Γ̂(z) =
1 − 2z −1 + z −2 Γ(z) 1 − 2z −1 + z −2 + (α + β)z −(d+1) − αz −(d+2)
㹷ӹͧà´ÕÂǡѹ ࡳ±ì·Õ ÍÒ¨¨Ð¹ÓÁÒãªé㹡ÒÃàÅ×Í¡¤èÒ
β
¤×Í ¡ÒÃàÅ×Í¡¤èÒ
β
(2.15)
·Õ ·ÓãËéàʶÕÂÃÀÒ¾
¢Í§ÃкºáÅÐÍѵÃÒ¡ÒÃÅÙèà¢éÒà» ¹·Õ ¹èÒ¾Í㨠ÊÓËÃѺ»ÃÔÁÒ³¢Í§ÍͿ૵·Ò§¤ÇÒÁ¶Õ ·Õ ¡Ó˹´ÁÒãËé¢Í§
¡ÒÃÍÍ¡à຺¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL
Maximum magnitude of E(z) after C samples
2.3.
0.2
C = 100
0.18
C = 50 0.16
0.5% frequency offset
0.14 0.12 0.1 0.08 0.06 0.04 0.02 0
0.2% frequency offset 0
0.5
ÃÙ»·Õ 2.6: ¢¹Ò´ÁÒ¡ÊØ´¢Í§
áµèÅÐ
d, C
áÅÐ
29
αC
E(z)
1
β (PLL gain parameter)
ËÅѧ¨Ò¡·Õ ¢éÍÁÙżèÒ¹ä»
C
β
2 −3
x 10
ºÔµ àÁ× Íãªé
«Ö § ÊÒÁÒö·Õ ¨Ð¤Ó¹Ç³ËÒä´é ´Ñ§¹Õ ÊÓËÃѺ ¤èÒ
¢Ñ ¹µÍ¹áá ¤×Í ¡ÒÃàÅ×Í¡¤èÒ
1.5
d, C
d = 14
áÅÐ
αC
αC
áÅÐ
·Õ ¡Ó˹´ÁÒãËé
·Õ ·ÓãËéÃкºÁÕ¤ÇÒÁàʶÕÂÃÀÒ¾ ¨Ò¡ÊÁ¡Òà (2.14) áÅÐ (2.15) Ãкº
¨ÐÁÕ¤ÇÒÁàʶÕÂÃÀÒ¾¡çµèÍàÁ× Í ·Ø¡â¾ÅËÃ×ÍÃÒ¡¤ÓµÍº¢Í§µÑÇÊèǹ¢Í§ÊÁ¡Òà (2.14) ËÃ×Í (2.15) ÍÂÙè ÀÒÂã¹Ç§¡ÅÁË¹Ö §Ë¹èÇ áÅж֧áÁéÇèÒ¨ÐÁÕ¤èÒ
β
à¾Õ§¤èÒà´ÕÂÇ·Õ ·ÓãËé
E(z)
β
ËÅÒ¤èÒ·Õ ·ÓãËéÃкºÁÕ¤ÇÒÁàʶÕÂÃÀÒ¾ áµè¨ÐàÅ×Í¡
ÁÕ¢¹Ò´¹éÍÂÊØ´ ËÅѧ¨Ò¡·Õ ¢éÍÁÙżèÒ¹ä»
C
ºÔµ à¾× Í·Õ ¨ÐÅ´¼Å¡Ãзº
¢Í§ÊÑҳú¡Ç¹·Õ ¨Ðà¢éèÒÁÒã¹Ç§¨Ã PLL ÃÙ»·Õ 2.6 áÊ´§¢¹Ò´ÁÒ¡ÊØ´ (maximum magnitude) ¢Í§
E(z)
µÒÁÊÁ¡Òà (2.15) ËÅѧ¨Ò¡·Õ ¢éÍÁÙżèÒ¹ä»
¾ºÇèÒ ÇÔ¸Õ¡ÒÃÇÔà¤ÃÒÐËì¹Õ ¨ÐãËéä´é¤èÒ à¾Õ§áµè¢¹Ò´¢Í§
ËÁÒÂà˵Ø
E(z)
β
C
ºÔµ ÊÓËÃѺ
d = 14T
áÅÐ
αC
Êѧࡵ¨Ð
à» ¹¤èÒà´ÕÂǡѹ â´ÂäÁè¤Ó¹Ö§¶Ö§»ÃÔÁÒ³¢Í§ÍͿ૵·Ò§¤ÇÒÁ¶Õ
¨ÐµèÒ§¡Ñ¹à·èÒ¹Ñ ¹ µÒÁ»ÃÔÁÒ³¢Í§ÍͿ૵·Ò§¤ÇÒÁ¶Õ
ÇÔ¸Õ¡ÒÃÍ͡Ẻ¤èÒ ¾ÒÃÒÁÔàµÍÃì ¢Í§Ç§¨Ã PLL µÒÁ·Õ ¡ÅèÒÇÁÒ¢éÒ§µé¹ ¹Õ ¨ÐÍÂÙè º¹¾× ¹°Ò¹
¢Í§ÊÁÁص԰ҹ·Õ ÇèÒ ¤ÇÒÁªÑ¹ ¢Í§àÊé¹â¤é§ ÃÙ» µÑÇ àÍÊ (S curve) ¢Í§ÊÁ¡Òà (2.4) ÁÕ ¤èÒ à» ¹ ¤èÒ Ë¹Ö § ³ ¨Ø´¡Óà¹Ô´ áÅÐäÁè ÁÕ ÊÑҳú¡Ç¹ÀÒÂã¹Ç§¨Ã TED à¾ÃÒÐ©Ð¹Ñ ¹ ¡è͹·Õ ¨Ð¹Ó¤èÒ
α
áÅÐ
β
·Õ ä´é
30
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¨Ò¡ÇÔ¸Õ¡ÒÃÍ͡ẺµÒÁ·Õ ¡ÅèÒÇÁÒ¢éÒ§µé¹ ¹Õ ä»ãªé §Ò¹ ¼Ùéãªé ¨Ðµéͧ·ÓãËé ¤ÇÒÁªÑ¹ ¢Í§àÊé¹â¤é§ ÃÙ» µÑÇ àÍÊ ¢Í§Ç§¨Ã TED ÁÕ¤èÒà» ¹¤èÒË¹Ö § ³ ¨Ø´¡Óà¹Ô´¡è͹àÊÁÍ
2.3.3
¡ÒÃËÒàÊé¹â¤é§ÃÙ»µÑÇàÍÊ
ä·ÁÁÔ § ¿ §¡ìªÑ¹ (timing function) ËÃ×Í àÊé¹â¤é§ ÃÙ» µÑÇ àÍÊ (S curve) [22] ¨Ð¹ÔÂÒÁâ´Â ¤èÒà©ÅÕ Â (mean) ¢Í§ =
rk
{²̂k }â´ÂÊÁÁصÔÇèÒ
ÊÓËÃѺ·Ø¡¤èÒ
k,
¤èÒ
r̂k
·Õ ä´é¨Ò¡Ç§¨ÃµÃǨËÒÊÑÅѡɳìÁÕ¤èÒ¶Ù¡µéͧ·Ñ §ËÁ´ ¹Ñ ¹¤×Í
r̂k
áÅТéÍÁÙÅÍÔ¹¾ØµáµèÅкԵäÁèÁÕÊËÊÑÁ¾Ñ¹¸ì¡Ñ¹ (uncorrelated) áÅÐÁÕ¾Åѧ§Ò¹
à·èҡѺ 1 ˹èÇ ´Ñ§¹Ñ ¹
STED (²) = E[²̂k | ², r̂k = rk for ∀k] àÁ× Í
² = τ − τ̂
(2.16)
¤×Í ¢éͼԴ¾ÅÒ´·Ò§àÇÅÒ à¹× ͧ¨Ò¡ ÃÙ» ¡ÃÒ¿¢Í§ä·ÁÁÔ § ¿ §¡ìªÑ¹ ÁÕ ÅѡɳФÅéÒÂ
µÑÇÍÑ¡Éà S (àÁ× ÍËÁعÃÙ»¡ÃÒ¿ 90 ͧÈÒ) ´Ñ§¹Ñ ¹ ¨Ö§àÃÕ¡¡Ñ¹ÇèÒ àÊé¹â¤é§ÃÙ»µÑÇàÍÊ (S curve) «Ö § ÊÒÁÒö·Õ ¨Ð¹ÓÁÒãªéÇÑ´»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ç§¨Ã TED ä´é ¹Í¡¨Ò¡¹Õ ã¹¡Ã³Õ ·Õ ·ÃÒºÇèÒ ¼ÅµÍºÊ¹Í§ÍÔÁ¾ÑÅÊì¢Í§ªèͧÊÑÒ³¤×Í ÍÐäà àÊé¹â¤é§ ÃÙ» µÑÇ àÍÊ ÊÒÁÒö·Õ ¨Ð¤Ó¹Ç³ËÒä´é â´ÂµÃ§ [22] µÑÇÍÂèÒ§àªè¹ ãËé ¾Ô¨ÒóÒẺ¨ÓÅͧ¢Í§ªèͧÊÑÒ³ PR4 (partial response class IV) ã¹ÃÙ»·Õ 2.2 àÁ× Í
H(D)
1 − D2
=
´Ñ§¹Ñ ¹ àÊé¹â¤é§ÃÙ»µÑÇàÍʢͧǧ¨Ã
M&M TED ÊÓËÃѺªèͧÊÑÒ³¹Õ ËÒä´é¨Ò¡
STED (²) = E[²̂k | ², r̂k−1 = rk−1 , r̂k = rk ] = KT E[rk−1 yk − rk yk−1 ] X = KT E[(ak−1 − ak−3 ) ai h(kT − iT − ²) i
− (ak − ak−2 )
X
ai h(kT − T − iT − ²)]
i
= â´Â·Õ
P
rk = ak −ak−2
i ai h(kT
− iT − ²)
3T {−h(−T − ²) + 2h(T − ²) − h(3T − ²)} 16
(2.17)
¤×Í ¢éÍÁÙÅàÍÒµì¾ØµµÑÇ·Õ
k
¢Í§ªèͧÊÑÒ³·Õ »ÃÒȨҡÊÑҳú¡Ç¹,
¤×Í ¢éÍÁÙÅàÍÒµì¾ØµµÑÇ·Õ
k
¢Í§Ç§¨ÃªÑ¡µÑÇÍÂèÒ§, áÅÐ
h(t)
=
yk
=
q(t) − q(t − 2T )
2.3.
¡ÒÃÍÍ¡à຺¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL
31
n(t) PR-IV pulse
ak
h(t) = q(t) - q(t-2T)
τ
p(t)
LPF
y( t )
yk t k = kT
to data detection
z −d
symbol detector
yk −d
rˆk −d
εˆk
TED
ÃÙ»·Õ 2.7: Ẻ¨ÓÅͧÊÓËÃѺ¡ÒÃËÒàÊé¹â¤é§ÃÙ»µÑÇàÍÊ
¤×Í ÊÑÒ³¾ÑÅÊìẺ PR4 ¤èÒ¤§µÑÇ
KT
=
3T /16
¶Ù¡ãªéà¾× ͪèÇ·ÓãËé¤ÇÒÁªÑ¹¢Í§àÊé¹â¤é§ÃÙ»µÑÇ
àÍÊã¹ÊÁ¡Òà (2.17) ÁÕ¤èÒà·èҡѺ¤èÒË¹Ö § ³ ¨Ø´¡Óà¹Ô´
ÊÓËÃÑºã¹¡Ã³Õ·Õ äÁè·ÃÒº¼ÅµÍºÊ¹Í§ÍÔÁ¾ÑÅÊì¢Í§ªèͧÊÑÒ³ àÊé¹â¤é§ÃÙ»µÑÇàÍÊ¡çÂѧÊÒÁÒö·Õ ¨ÐËÒä´éâ´Â¡Ò÷ӡÒèÓÅͧÃкº «Ö §·Óä´éâ´Â¡ÒÃà» ´ä·ÁÁÔ §ÅÙ»¢Í§áºº¨ÓÅͧã¹ÃÙ»·Õ 2.2 (¡ÅèÒǤ×Í µÑ´Ç§¨Ã¡ÃͧÅÙ»áÅÐǧ¨Ã VCO ÍÍ¡¨Ò¡áºº¨ÓÅͧ) «Ö §¨Ð·ÓãËéä´éà» ¹áºº¨ÓÅͧãËÁèµÒÁÃÙ»·Õ 2.7 ¨Ò¡ÃÙ» ÊÑÒ³
τ
´éÇÂ
²
¨Ð¶Ù¡ ·Ó¡Òêѡ µÑÇÍÂèÒ§·Õ àÇÅÒ
kT
(¹Ñ ¹¤×Í ¡Ó˹´ãËé
¨Ò¡¹Ñ ¹ ·Ó¡ÒäӹdzËÒ¤èÒà©ÅÕ Â ·Ò§àÇÅÒ (time average) ¢Í§
à¾× ÍãËéä´éà» ¹¤èÒ
−0.5T
y(t)
¶Ö§¤èÒ
STED (²)
0.5T
¤èÒà´ÕÂÇ ·ÓÅÑ¡É³Ð¹Õ ä»àÃ× ÍÂæ à¾× ÍËÒ¤èÒ
ÊØ´·éÒ¡ç·Ó¡ÒÃÇÒ´¡ÃÒ¿ÃÐËÇèÒ§
²/T
áÅÐ
τ̂ = 0)
{²̂k }
STED (²)
STED (²)/T
áÅÐá·¹¤èÒ
ÊÓËÃѺ áµèÅФèÒ
ÊÓËÃѺ
²
²
ÊÓËÃѺ¤èÒ
à¾× ÍãËéä´éà» ¹àÊé¹â¤é§
ÃÙ»µÑÇàÍÊ
ÃÙ» ·Õ 2.8 áÊ´§àÊé¹â¤é§ ÃÙ» µÑÇ àÍʢͧǧ¨Ã M&M TED ÊÓËÃѺ ªèͧÊÑҳẺ PR4 ·Õ ãªé ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº·Õ ãªé ¡Ñ¹ ·Ñ Çä» ³ ÃдѺ SNR µèÍ ºÔµ ËÃ×Í
Eb /N0
µèÒ§æ ÁÕ Ë¹èÇÂà» ¹ à´«ÔàºÅ
(dB) â´Âãªéǧ¨Ã PLL ¨Ðãªé¤èҵѴÊԹ㨢³ÐË¹Ö §áººá¢ç§ (instantaneous hard decision) ·Õ ä´éÃѺ ¨Ò¡Ç§¨ÃµÃǨËÒ¢Õ´àÃÔ Áà»ÅÕ Â¹·Õ äÁèÁÕ˹èǤÇÒÁ¨Ó (memoryless threshold detector) Ẻ 3 ÃдѺ
32
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
0.5 0.4 0.3
10 dB
S TED(ε)/T
0.2
20 dB 0.1
Mean 0
Eb/N0 = 5 dB −0.1 −0.2 −0.3 −0.4
Normalized timing funtion −0.5 −0.5
0
0.5
Normalized timing error (ε/T)
ÃÙ»·Õ 2.8: àÊé¹â¤é§ÃÙ»µÑÇàÍʢͧǧ¨Ã M&M TED ÊÓËÃѺªèͧÊÑÒ³ PR4 ·Õ ãªéä·ÁÁÔ §ÃԤѿàÇÍÃÕ áºº·Õ ãªé¡Ñ¹·Ñ Çä»
â´ÂÁÕÃдѺ¢Õ´àÃÔ Áà»ÅÕ Â¹ (threshold level) ·Õ ¤èÒ
r̂k =
±1
¹Ñ ¹¤×Í
2 if yk > 1 −2 if yk < −1
(2.18)
0 else
àÊ鹡ÃÒ¿¢Í§ ä·ÁÁÔ §¿ §¡ìªÑ¹áºº¹ÍÃìÁÍÅäÅ«ì (normalized timing function) ¨Ðä´éÁÒ¨Ò¡ÊÁ¡Òà (2.17) ¨Ò¡ÃÙ»·Õ 2.8 ¨Ð¾ºÇèÒ àÊé¹â¤é§ÃÙ»µÑÇàÍÊÁÕÅѡɳÐÊÁÁÒµÃẺ¤Õ (odd symmetric) àÁ× Íà·Õº ¡Ñº
²=0
«Ö §ËÁÒ¤ÇÒÁËÁÒÂÇèÒ ÍͿ૵·Ò§à¿Ê¢Í§¡ÒêѡµÑÇÍÂèÒ§·Õ ¶Ù¡»ÃѺ¤èÒ´éÇÂǧ¨Ã PLL ¨Ð
ÊÔ ¹ÊØ´ ³ ¨Ø´àʶÕÂÃÀÒ¾ (stable point) ·Õ
²=0
Êѧࡵ¨Ð¾ºÇèÒ àÊé¹â¤é§ÃÙ»µÑÇàÍÊ
ä´é¨Ò¡¡ÒèÓÅͧÃкº¨ÐÊÍ´¤Åéͧ¡Ñºä·ÁÁÔ §¿ §¡ìªÑ¹áºº¹ÍÃìÁÍÅäÅ«ìàÁ× Í à¾ÃÒÐÇèÒ ÊÁÁص԰ҹ·Õ ¡Ó˹´ãËé
r̂k
à» ¹ à˵ؼÅÇèÒ ·ÓäÁªèǧ¢Í§¡ÃÒ¿·Õ
=
rk
ÊÓËÃѺ·Ø¡¤èÒ
STED (²)/T
k
¨Ðãªéä´éäÁè´Õ àÁ× Í
²/T ²/T
STED (²)/T
·Õ
ÁÕ¤èÒ¹éÍ ·Ñ §¹Õ à» ¹ ÁÕ¤èÒÁÒ¡ ´Ñ§¹Ñ ¹¨Ö§
ÊÍ´¤Åéͧ¡Ñº ä·ÁÁÔ § ¿ §¡ìªÑ¹ Ẻ¹ÍÃìÁÍÅäÅ«ì àÁ× Í
2.3.
¡ÒÃÍÍ¡à຺¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL
33
T
Timing estimate
0.5T
Estimated τ 0
T
−0.5T
Actual τ −T
0
500
1000
1500
2000
2500
3000
3500
4000
Time (in bit periods)
ÃÙ»·Õ 2.9: µÑÇÍÂèÒ§ÅѡɳТͧä«à¤ÔÅÊÅÔ»
Eb /N0
ÁÕ¤èÒÊÙ§ ¨Ö§ÁÕ¤ÇÒÁ¡ÇéÒ§ÁÒ¡¡ÇèÒªèǧ¢Í§¡ÃÒ¿ àÁ× Í
¹Í¡¨Ò¡¹Õ ¨Ø´ ·Õ àÊé¹â¤é§ ÃÙ» µÑÇ àÍʵѴ ¡Ñº àÊé¹ á¡¹
x
Eb /N0
ÁÕ¤èÒ¹éÍÂ
¹Ñ ¹¤×Í àÁ× Í
STED (²)/T
0
=
¨ÐàÃÕ¡ÇèÒ
¨Ø´ÊÁ´ØÅ (equilibrium point) ¢Í§¡Ò÷ӧҹ «Ö §à» ¹¨Ø´·Õ ǧ¨Ã PLL ¨ÐÊÒÁÒöµÔ´µÒÁÍͿ૵ ·Ò§àÇÅÒä´é à» ¹ ÍÂèÒ§´Õ ÍÂèÒ§äáçµÒÁã¹·Ò§»¯ÔºÑµÔ ¨Ø´ ÊÁ´ØÅ ¨ÐÁÕ ËÅÒµÓáË¹è§ ¤×Í
±2T, . . . , ±nT
àÁ× Í
²
=
0, ±T ,
n ¤×Í àÅ¢¨Ó¹Ç¹àµçÁ à¹× ͧ¨Ò¡ ÊÑҳú¡Ç¹áÅСÒÃú¡Ç¹ (disturbance)
ã¹ÃкºÍÒ¨¨Ð·ÓãËé à¡Ô´ ¢éͼԴ¾ÅÒ´¢¹Ò´ãËè ÃÐËÇèÒ§¡Ãкǹ¡ÒûÃѺ ¤èÒ ÍͿ૵·Ò§à¿Ê¢Í§¡Òà ªÑ¡ µÑÇÍÂèÒ§ «Ö § à» ¹ ¼Å·ÓãËé¡Ò÷ӧҹ¢Í§Ç§¨Ã PLL à¡Ô´ ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð¨Ò¡¨Ø´ ÊÁ´ØÅ ¨Ø´ Ë¹Ö § ä» Âѧ ¨Ø´ ÊÁ´ØÅ ÍÕ¡ ¨Ø´ Ë¹Ö § à˵ءÒóì àªè¹¹Õ ¨ÐàÃÕ¡¡Ñ¹ ÇèÒ ä«à¤ÔÅÊÅÔ» (cycle slip) «Ö § ¨Ð·ÓãËé à¡Ô´ ¢éͼԴ¾ÅÒ´¨Ó¹Ç¹ÁÒ¡·Õ ǧ¨ÃµÃǨËÒ (detector) ÃÙ» ·Õ 2.9 áÊ´§µÑÇÍÂèÒ§ÅѡɳТͧä«à¤ÔÅÊÅÔ» ¨Ò¡ÃÙ»¨Ð¾ºÇèÒ Ç§¨Ã PLL ÊÒÁÒöµÔ´µÒÁ¡ÒÃà»ÅÕ Â¹á»Å§¢Í§
τ
ä´é´Õ ³ ¨Ø´àÃÔ Áµé¹¢Í§¡ÅØèÁ¢éÍÁÙÅ
(data packet) áµè àÁ× Í ÁÕ ä«à¤ÔÅÊÅÔ» à¡Ô´ ¢Ö ¹ ǧ¨Ã PLL ¨Ð¤èÍÂæ ÊÙàÊÕ ¡ÒõԴµÒÁ¤èÒ
τ
¨¹¡ÃÐ·Ñ §
ǧ¨Ã PLL à¢éÒ ÊÙè ¨Ø´ ÊÁ´ØÅ ãËÁè ÍÕ¡ ¨Ø´ Ë¹Ö § ´Ñ§¹Ñ ¹ ÍÒ¨¨Ð¡ÅèÒÇä´é ÇèÒ ä«à¤ÔÅÊÅÔ» à» ¹ ÊÒà赯 ·ÓãËé ǧ¨Ã
34
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
PLL ·Ó§Ò¹·Õ ¨Ø´ÊÁ´ØÅÍÕ¡¨Ø´Ë¹Ö § «Ö §à» ¹à˵ؼÅÇèÒ·ÓäÁ
τ̂
¨Ö§ÁÕ¤èÒµèÒ§¨Ò¡
τ
»ÃÐÁÒ³
T
àÁ× ÍÊÔ ¹ÊØ´
¢Í§¡ÅØèÁ ¢éÍÁÙÅ ¨ÐàËç¹ä´éÇèÒ ä«à¤ÔÅÊÅԻ໠¹ ÊÔ § ·Õ ÍѹµÃÒÂÁÒ¡ÊÓËÃѺ Ãкºä·ÁÁÔ § ÃԤѿàÇÍÃÕ ´Ñ§¹Ñ ¹ ¹Ñ¡ÇԨѠ¨Ö§ ä´é àʹÍÇÔ¸Õ ¡ÒõèÒ§æ ·Õ ¨Ð¹Óãªé 㹡ÒèѴ¡ÒáѺ ä«à¤ÔÅÊÅÔ» ÊÓËÃѺ ¼Ùé ʹã¨ÊÒÁÒöÈÖ¡ÉÒ ÃÒÂÅÐàÍÕ´ä´éã¹ [29, 30, 31, 32, 33]
2.4
»ÃÐÊÔ·¸ÔÀÒ¾¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕà຺·Õ ãªé¡Ñ¹·Ñ Çä»
¾Ô¨ÒóÒẺ¨ÓÅͧªèͧÊÑÒ³ PR4 ·Õ áÊ´§ã¹ÃÙ» ·Õ 2.2 ¨ÐàËç¹ä´éÇèÒ ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áÅÐǧ¨Ã µÃǨËÒÊÑÅÑ¡É³ì ¨Ð·Ó§Ò¹á¡¨Ò¡¡Ñ¹ ´Ñ§¹Ñ ¹ »ÃÐÊÔ·¸ÔÀÒ¾ÃÇÁ¢Í§Ãкº·Õ äÁè ä´é ¶Ù¡ à¢éÒ ÃËÑÊ (un coded system) ¨Ð¢Ö ¹ÍÂÙè¡Ñº¤Ø³ÀÒ¾¢Í§Ãкºä·ÁÁÔ §ÃԤѿàÇÍÃÕ ã¹Êèǹ¹Õ ¨ÐáÊ´§»ÃÐÊÔ·¸ÔÀÒ¾¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä» àÁ× Í·Ó§Ò¹ã¹Ãкº·Õ ÁÕáÅÐ äÁè ÁÕ ÍͿ૵·Ò§¤ÇÒÁ¶Õ ¹Í¡¨Ò¡¹Õ ǧ¨ÃµÃǨËÒÊÑÅÑ¡É³ì ·Õ ãªé ã¹Ç§¨Ã PLL ¤×Í Ç§¨ÃµÃǨËÒ ¢Õ´ àÃÔ Á à»ÅÕ Â¹·Õ äÁè ÁÕ Ë¹èǤÇÒÁ¨ÓẺËÅÒÂÃдѺ (multi level memoryless threshold detector) â´Â¢éÍÁÙÅ àÍÒµì¾Øµ ·Õ ä´é ¨Ðà» ¹ 仵ÒÁÊÁ¡Òà (2.18) ÊÓËÃѺ ¡Ãкǹ¡ÒõÃǨËÒ¢éÍÁÙÅ ÅӴѺ ¢éÍÁÙÅ
{yk }
¨Ð¶Ù¡Êè§ä»Âѧǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ·Õ ÁÕ»ÃÔÁҳ˹èǧàÇÅÒÊÓËÃѺ¡ÒõѴÊÔ¹ã¨à·èҡѺ
60T
à¾× Í
ËÒÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ ·Õ à» ¹ ä»ä´é ÁÒ¡·Õ ÊØ´ ¹Í¡¨Ò¡¹Õ BER áµèÅФèÒ ·Õ ä´é ¨Ð¶Ù¡ ¤Ó¹Ç³â´Âãªé ¡ÅØèÁ ¢éÍÁÙŨӹǹÁÒ¡¨¹¡ÇèÒ¨Ðà¡Ô´¢éͼԴ¾ÅÒ´ÃÇÁ·Ñ §ËÁ´ 1000 ºÔµ ÊÓËÃѺ Ãкº·Õ äÁè ÁÕ ÍͿ૵·Ò§¤ÇÒÁ¶Õ ǧ¨ÃÀÒ¤ÃѺ ÊÒÁÒö·Õ ¨Ðãªé ǧ¨Ã PLL Íѹ´Ñº ·Õ Ë¹Ö § ä´é ã¹¡Ã³Õ¹Õ ¨ÐÊÁÁصÔãËé ÃкºÁÕ¡ÒÃà¢éҨѧËÇÐ㹪èǧÀÒÇСÒÃä´éÁÒẺÊÁºÙóì (perfect acquisition) «Ö § ·Óä´é â´Â¡ÒáÓ˹´ãËé
τ0 = 0
à¾× Í ·Õ ÇèÒ ¨Ðä´é äÁè µéͧãªé ¢éÍÁÙÅ preamble áÅСÓ˹´ãËé ¢éÍÁÙÅ
Ë¹Ö § ¡ÅØèÁ ÁÕ ¨Ó¹Ç¹ 4096 ºÔµ ÃÙ» ·Õ 2.10 à»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§¢éͼԴ¾ÅÒ´·Ò§àÇÅÒẺÃÒ¡ ¡ÓÅѧ Êͧà©ÅÕ Â (RMS: root mean square) ¹Ñ ¹¤×Í
σ²
p
=
E[(τk − τ̂k )2 ]
áÅÐ BER â´Â¤èÒ
¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL (α) ·Õ ãªé¨Ð¶Ù¡Í͡ẺÁÒà¾× ÍãËéÊÒÁÒöµÔ´µÒÁ¡ÒÃà»ÅÕ Â¹á»Å§·Ò§à¿Ê ä´é ÀÒÂã¹ 100 ºÔµ (¹Ñ ¹¤×Í
α100
= 0.0295) µÒÁ·Õ ä´é ͸ԺÒÂã¹ËÑÇ¢éÍ ·Õ 2.3.1 ¨ÐàËç¹ä´éÇèÒ àÁ× Í
ÃдѺ ¤ÇÒÁÃعáç¢Í§¨Ôµ àµÍÃì ·Ò§àÇÅÒ ¢Í§
σ² /T
σ² /T
σw /T
áÅÐ BER) ¹Í¡¨Ò¡¹Õ ¶éÒÃкºÁÕ¤èÒ
ÂÔ § ÁÒ¡ »ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº¡ç ¨ÐÂÔ § áÂè (·Ñ § ã¹ÃÙ»
σ² /T
¹éÍ Ãкº¡ç¨ÐÁÕ BER µ Ó ´Ñ§¹Ñ ¹ ¾ÒÃÒÁÔàµÍÃì
áÅÐ BER ¨Ö§ÊÒÁÒö·Õ ¨Ð¹ÓÁÒãªéà» ¹à¡³±ì㹡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкºä·ÁÁÔ §
2.4.
»ÃÐÊÔ·¸ÔÀÒ¾¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕà຺·Õ ãªé¡Ñ¹·Ñ Çä»
35
18
RMS timing error σε/T (%)
16 14 12 10
σw/T = 1%
8 6 4
σ /T = 0.1% w
2 0
4
5
6
7
8
9
10
9
10
(a) Eb/N0 (dB) −1
10
−2
BER
10
σw/T = 1% σw/T = 0.1%
−3
10
−4
10
−5
10
4
5
6
7
8
(b) Eb/N0 (dB)
ÃÙ»·Õ 2.10:
(a) ¢éͼԴ¾ÅÒ´·Ò§àÇÅÒẺ RMS
ªèͧÊÑÒ³ÍØ´Á¤µÔẺ PR4 ·Õ ÁÕ¤èÒ
σw /T
σ² /T
áÅÐ (b) »ÃÐÊÔ·¸ÔÀÒ¾¢Í§ BER ÊÓËÃѺ
µèÒ§æ ¡Ñ¹ (àÁ× ÍÃкºäÁèÁÕÍͿ૵·Ò§¤ÇÒÁ¶Õ )
ÃԤѿàÇÍÃÕẺµèÒ§æ ä´é 㹷ӹͧà´ÕÂǡѹ ÊÓËÃѺÃкº·Õ ÁÕÍͿ૵·Ò§¤ÇÒÁ¶Õ ǧ¨ÃÀÒ¤ÃѺ¨Ðµéͧãªéǧ¨Ã PLL Íѹ´Ñº·Õ
36
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
0
10
Conventional timing recovery with hard decision −1
10
C = 50 C = 100
−2
10
BER
C = 256 Perfect timing
−3
10
−4
10
−5
10
5
6
7
8
9
10
Eb/N0 (dB)
ÃÙ»·Õ 2.11: »ÃÐÊÔ·¸ÔÀÒ¾ BER ¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Ç仢ͧªèͧÊÑÒ³ÍØ´Á¤µÔẺ PR4 ÊÓËÃѺ
σw /T = 0.5%
áÅÐÍͿ૵·Ò§¤ÇÒÁ¶Õ 0.2%
Êͧ㹡ÒèѴ¡ÒáѺÍͿ૵·Ò§¤ÇÒÁ¶Õ ¹Õ â´Â¨Ð¾Ô¨ÒóÒÃкº·Õ ·Ó§Ò¹ã¹ÊÀÒÇлҹ¡ÅÒ§ (moder ate condition) ¹Ñ ¹¤×ÍÁÕ áÅÐ
C
β
σw /T = 0.5%
áÅÐÍͿ૵·Ò§¤ÇÒÁ¶Õ 0.2% àªè¹à´ÕÂǡѹ ¾ÒÃÒÁÔàµÍÃì
·Õ ãªé¨Ð¶Ù¡Í͡ẺÁÒà¾× ÍãËéÊÒÁÒöµÔ´µÒÁ¡ÒÃà»ÅÕ Â¹á»Å§·Ò§à¿ÊáÅзҧ¤ÇÒÁ¶Õ ä´éÀÒÂã¹
ºÔµ µÒÁ·Õ ä´é ͸ԺÒÂã¹ËÑÇ¢éÍ ·Õ 2.3.2 «Ö § ¨Ðä´é ÇèÒ ¤èÒ
α
·Õ ¶Ù¡ Í͡ẺÊÓËÃѺ
= 50, 100, áÅÐ 256 ¤×Í 0.012, 0.029, áÅÐ 0.058, µÒÁÅӴѺ ã¹¢³Ð·Õ ¤èÒ ÊÓËÃѺ
α
d=0
áÅÐ
C
d = 0 β
áÅÐ
C
·Õ ¶Ù¡ Í͡Ẻ
= 50, 100, áÅÐ 256 ¤×Í 0.00015, 0.000885, áÅÐ 0.00325, µÒÁÅӴѺ
¹Í¡¨Ò¡¹Õ ¨Ð¾Ô¨ÒóÒ੾ÒÐ¡Ã³Õ·Õ Ç§¨Ã PLL ãªé¤èÒ
α
áÅФèÒ
β
à´ÕÂǡѹ ÊÓËÃѺ·Ñ §ªèǧÀÒÇСÒÃä´é
ÁÒáÅÐÀÒÇСÒõԴµÒÁ â´Â¢éÍÁÙÅË¹Ö §¡ÅØèÁ¨Ð»ÃСͺ仴éÇ preamble ¨Ó¹Ç¹
C
ºÔµ áÅÐÊèǹ·Õ
à» ¹ºÔµ¢èÒÇÊÒèӹǹ 4096 ºÔµ ÃÙ»·Õ 2.11 áÊ´§»ÃÐÊÔ·¸ÔÀÒ¾¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä» àÁ× Íãªé¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL ·Õ Í͡ẺÊÓËÃѺáµèÅÐ ãªé á·¹ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº·Õ ãªé ¡Ñ¹ ·Ñ Çä» ·Õ ãªé
τ̂k
=
τk
C
àÊ鹡ÃÒ¿·Õ à¢Õ¹ÇèÒ Perfect timing
ÊÓËÃѺ ¡Òêѡ µÑÇÍÂèÒ§ÊÑÒ³
y(t)
¨Ð
2.5.
ä·ÁÁÔ §ÃԤѿàÇÍÃÕà຺´Ô¨Ô·ÑÅ
37
àËç¹ä´éÇèÒ ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä»äÁèÊÒÁÒö·Ó§Ò¹ä´é´Õ àÁ× Í·Ó§Ò¹ã¹Ãкº·Õ µéͧ¡ÒÃÍѵÃÒ ¡ÒÃÅÙèà¢éÒ·Õ ÃÇ´àÃçÇ ËÃ×Í ÍÕ¡ ¹ÑÂ Ë¹Ö § ¡ç ¤×Í àÁ× Í ãªé §Ò¹¡Ñº ¤èÒ ¾ÒÃÒÁÔàµÍÃì ¢Í§Ç§¨Ã PLL ·Õ ¶Ù¡ Í͡Ẻ ÊÓËÃѺ
C
2.5
ä·ÁÁÔ §ÃԤѿàÇÍÃÕà຺´Ô¨Ô·ÑÅ
¹éÍÂæ
ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº·Õ ãªé ¡Ñ¹ ·Ñ Çä»·Õ Í¸ÔºÒÂ㹺·¹Õ ¨ÐÁÕ ÅѡɳСÒ÷ӧҹ໠¹ Ẻ¼ÊÁ (hybrid) ¹Ñ ¹¤×Í ÁÕ ·Ñ § Êèǹ·Õ ·Ó§Ò¹¡Ñº ÊÑÒ³á͹ÐÅçÍ¡ áÅÐÊèǹ·Õ ·Ó§Ò¹¡Ñº ÊÑÒ³´Ô¨Ô·ÑÅ ¾Ô¨ÒóҨҡ ÃÙ» ·Õ 2.2 â´Â·Ñ Çä» Ç§¨Ã VCO ÁÑ¡¨Ðà» ¹ ǧ¨Ãá͹ÐÅçÍ¡«Ö § ÁÕ ÅѡɳСÒ÷ӧҹ·Õ «Ñº«é͹áÅÐÁÕ ÃÒ¤Òᾧ ã¹Êèǹ¹Õ ¨Ð¡ÅèÒǶ֧Ãкºä·ÁÁÔ §ÃԤѿàÇÍÃÕ·Õ ÁÕÅѡɳÐà» ¹áºº´Ô¨Ô·ÑÅ·Ñ §ËÁ´ «Ö §ÁÕãªé㹪Ի ªèͧÊÑÒ³ÍèÒ¹ (read channel chip) ºÒ§ÃØè¹ ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº´Ô¨Ô·ÑÅ ¨Ðãªé ÍѵÃÒ¡Òêѡ µÑÇÍÂèÒ§ (sampling rate) ·Õ äÁè à¢éÒ ¨Ñ§ËÇÐ (asyn chronous) ¡ÑºÊÑÒ³á͹ÐÅçÍ¡·Õ ä´éÃѺ à¾Õ§áµè¢ÍãËéÁÕ¤ÇÒÁ¶Õ ¡ÒêѡµÑÇÍÂèÒ§ (sampling frequen 5
cy) ÊÙ§¡ÇèÒ¤ÇÒÁ¶Õ 乤ÇÔµÊì
(Nyquist frequency) [2, 10] ¢Í§ÊÑÒ³á͹ÐÅçÍ¡ ËÃ×ÍÍÒ¨¨Ð¡ÅèÒÇ 6
ä´éÇèÒ Ç§¨ÃªÑ¡µÑÇÍÂèÒ§ãªéÍѵÃÒ¡ÒêѡµÑÇÍÂèҧẺà¡Ô¹¨ÃÔ§
(oversampling rate) [34, 35] à¹× ͧ¨Ò¡
¢éÍÁÙÅá«Áà» Å·Õ ä´é¨Ò¡Ç§¨ÃªÑ¡µÑÇÍÂèÒ§¹Õ ¨ÐäÁèà¢éҨѧËÇСѺ¢éÍÁÙźԵ·Õ Êè§ÁÒ¨Ò¡µé¹·Ò§ ´Ñ§¹Ñ ¹ ¨Ö§µéͧ ÁÕ ¡ÒûÃѺ¤èÒ·Ò§àÇÅÒ (timing adjustment) ´éÇÂÇÔ¸Õ¡Ò÷ҧ´Ô¨Ô·ÑÅ·Õ àÃÕ¡ÇèÒ à·¤¹Ô¤¡ÒûÃÐÁÒ³ ¤èÒ ã¹ªèǧ (interpolation technique) à¾× ÍãËé ä´é ¢éÍÁÙÅ á«Áà» Å ·Õ ÊÍ´¤Åéͧ¡Ñº ¢éÍÁÙÅ ºÔµ ·Õ Êè§ ÁÒ¨Ò¡ µé¹·Ò§ Ãкºä·ÁÁÔ § ÃԤѿàÇÍÃÕ ·Õ ãªé à·¤¹Ô¤ ¹Õ ¨ÐàÃÕ¡¡Ñ¹ ·Ñ Çä»ÇèÒ ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº»ÃÐÁÒ³¤èÒ ã¹ªèǧ (interpolated timing recovery) «Ö § ÁÕ â¤Ã§ÊÃéÒ§µÒÁÃÙ» ·Õ 2.12 â´ÂÊèǹ»ÃСͺ·Ø¡ Êèǹ ã¹Ç§¨Ã PLL ¨ÐÁÕ ÅѡɳСÒ÷ӧҹ໠¹ Ẻ´Ô¨Ô·ÑÅ ·Ñ §ËÁ´ «Ö § ¨ÐªèÇ·ÓãËé ÊÒÁÒöŴ¤èÒãªé¨èÒÂã¹ ¡ÒÃÊÃéÒ§Ãкºä·ÁÁÔ § ÃԤѿàÇÍÃÕ ä´é ¨Ò¡¡Ò÷´Åͧ¾ºÇèÒ ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº»ÃÐÁÒ³¤èÒ ã¹ªèǧÁÕ »ÃÐÊÔ·¸ÔÀÒ¾à·Õºà·èÒ ¡Ñº ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº·Õ ãªé ¡Ñ¹ ·Ñ Çä» ¶éÒ ÁÕ ¡ÒÃàÅ×Í¡ãªé ǧ¨Ã¡Ãͧ¡ÒûÃÐÁÒ³ ¤èÒ ã¹ ªèǧ (interpolation lter) ·Õ àËÁÒÐÊÁ ÊÓËÃѺ ¼Ùé ʹ㨠ÊÒÁÒö ÈÖ¡ÉÒ ÃÒÂÅÐàÍÕ´ ¢Í§ ä·ÁÁÔ § ÃԤѿàÇÍÃÕẺ»ÃÐÁÒ³¤èÒ㹪èǧà¾Ô ÁàµÔÁä´é¨Ò¡ [34, 35, 36, 37] 5
¤ÇÒÁ¶Õ 乤ÇÔµÊì¢Í§ÊÑÒ³á͹ÐÅçÍ¡ ÁÕ¤èÒà·èҡѺÊͧà·èҢͧ¤ÇÒÁ¶Õ ÊÙ§ÊØ´¢Í§ÊÑÒ³á͹ÐÅçÍ¡¹Ñ ¹
6
¨Ó¹Ç¹¢éÍÁÙÅá«Áà» Å·Õ ä´é¨Ò¡Ç§¨ÃªÑ¡µÑÇÍÂèÒ§ ¨ÐÁըӹǹÁÒ¡¡ÇèÒ¢éÍÁÙÅÍÔ¹¾ØµºÔµ·Õ Êè§ÁҨҡǧ¨ÃÀÒ¤Êè§
38
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
A/D converter
y (t )
interpolation filter
t i = iTs
yk
Viterbi detector
âk
t k = ( mk + µ k )Ts
fixed sampling frequency (1 / Ts )
τˆk digital accumulator
symbol detector
z −d
interpolator control unit
loop filter
εˆk
yk − d
rˆk −d
TED
ÃÙ»·Õ 2.12: â¤Ã§ÊÃéÒ§¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ»ÃÐÁÒ³¤èÒ㹪èǧ
2.6
àà¹Çâ¹éÁ¢Í§Ãкºä·ÁÁÔ §ÃԤѿàÇÍÃÕã¹Í¹Ò¤µ
¨Ò¡¼ÅÅѾ¸ì ·Õ áÊ´§ã¹ÃÙ» ·Õ 2.10 áÅÐ 2.11 ÊÃØ» ä´é ÇèÒ ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº·Õ ãªé ¡Ñ¹ ·Ñ Ç仨зӧҹ ä´é äÁè ´Õ ¶éÒ ÃкºÁÕ ¢éͼԴ¾ÅÒ´·Ò§àÇÅÒÁÒ¡ ËÃ×Í àÁ× Í ·Ó§Ò¹ã¹Ãкº·Õ µéͧ¡ÒÃÍѵÃÒ¡ÒÃÅÙèà¢éÒ·Õ ÃÇ´àÃçÇ ÇÔ¸Õ¡ÒÃá¡é䢻 ËÒ·Õ §èÒÂ·Õ ÊØ´ 㹡ÒÃà¾Ô Á »ÃÐÊÔ·¸ÔÀÒ¾¢Í§ä·ÁÁÔ § ÃÔ ¤Ñ¿ àÇÍÃÕ áºº·Õ ãªé ¡Ñ¹ ·Ñ Çä» ¡ç ¤×Í ¡ÒÃà»ÅÕ Â¹Ç§¨ÃµÃǨËÒÊÑÅÑ¡É³ì ·Õ ãªé ã¹ä·ÁÁÔ § ÅÙ» ¨Ò¡ ǧ¨ÃµÃǨËÒ¢Õ´ àÃÔ Á à»ÅÕ Â¹áººá¢ç§ (hard threshold detector) ä»à» ¹Ç§¨ÃµÃǨËÒ¢Õ´àÃÔ Áà»ÅÕ Â¹áººÍè͹ (soft threshold detector) [33] ËÃ×Í ÍÒ¨¨Ðãªé ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ·Õ ÁÕ »ÃÔÁҳ˹èǧàÇÅÒÊÓËÃѺ ¡ÒõѴÊÔ¹ã¨
dT
ÊÑ ¹ ¡çä´é [27] ÍÂèÒ§äÃ
¡çµÒÁ ÇÔ¸Õ¡ÒÃ·Õ ¡ÅèÒÇÁÒàËÅèÒ¹Õ ¨ÐªèÇÂà¾Ô Á»ÃÐÊÔ·¸ÔÀÒ¾·Õ ä´éà¾Õ§àÅ硹éÍÂà·èÒ¹Ñ ¹ ´Ñ§¹Ñ ¹ Ãкºä·ÁÁÔ §ÃԤѿàÇÍÃÕẺãËÁè·Õ ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡¡ÇèÒä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä» ¨Ö§à» ¹ÊÔ §·Õ µéͧ¡ÒÃÍÂèÒ§ÁÒ¡ ã¹ [30, 38] ä´é¹ÓàʹÍä·ÁÁÔ §ÃԤѿàÇÍÃÕÃٻẺãËÁè·Õ àÃÕ¡ÇèÒ à¾Íà«ÍÃì äÇàÇÍÃìä·ÁÁÔ §ÃԤѿàÇÍÃÕ (per survivor timing recovery) «Ö §ãªé§Ò¹¡ÑºÃкº·Õ äÁèä´é¶Ù¡à¢éÒÃËÑÊ «Ö §ÁÕ
y
¨ÐºÍ¡
10−4
ËÃ×ÍÍÕ¡
»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä» ´Ñ§áÊ´§ã¹ÃÙ»·Õ 2.13 â´Â·Õ àÊé¹á¡¹ ¶Ö§»ÃÔÁÒ³
Eb /N0
(ÁÕ˹èÇÂà» ¹ dB) ·Õ Ãкºµéͧ¡Òà 㹡ÒÃ·Õ ¨Ð·ÓãËéÃкºÁÕ BER =
¹ÑÂË¹Ö §¡ç¤×Í àÊé¹á¡¹
y
¨ÐºÍ¡¶Ö§¡ÓÅѧ (power) ·Õ ǧ¨ÃÀÒ¤Ê觵éͧãªé㹡ÒÃÊ觢éÍÁÙÅ à¾× Í·Õ ¨Ð·ÓãËé
àà¹Çâ¹éÁ¢Í§Ãкºä·ÁÁÔ §ÃԤѿàÇÍÃÕã¹Í¹Ò¤µ
39
11 Conventional timing recovery with hard decision (d = 0) Conventional timing recovery with tentative decision (d = 4) Per−survivor timing recovery (d = 0) Genie−aided detector (d = 0)
10.8 10.6
Eb/N0 required to achieve BER = 10
−4
(in dB)
2.6.
10.4 10.2 10 9.8 9.6 9.4 9.2 9 0.1
0.2
0.3
0.4
0.5
0.6
σw/T (%)
0.7
0.8
0.9
1
ÃÙ»·Õ 2.13: »ÃÐÊÔ·¸ÔÀÒ¾¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺµèÒ§æ ¢Í§ªèͧÊÑÒ³ÍØ´Á¤µÔẺ PR4
ǧ¨ÃÀÒ¤ÃѺÁÕ BER =
10−4
¨Ò¡ÃÙ»·Õ 2.13 àÊ鹡ÃÒ¿·Õ à¢Õ¹ÇèÒ Genie aided detector ËÁÒ¶֧
ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä»·Õ Ç§¨Ã PLL ãªé
r̂k
=
rk
(´ÙÃÙ»·Õ 2.2) 㹡ÒûÃѺ¤èÒÍͿ૵·Ò§
à¿Ê¢Í§¡ÒêѡµÑÇÍÂèÒ§µÑǶѴä», áÅÐ tentative decision (d = 4) ËÁÒ¶֧ ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ ·Õ ãªé ¡Ñ¹ ·Ñ Çä»·Õ Ç§¨ÃµÃǨËÒÊÑÅÑ¡É³ì ·Õ ãªéè ã¹ä·ÁÁÔ § ÅÙ» ¤×Í Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ·Õ ÁÕ »ÃÔÁҳ˹èǧ àÇÅÒ
d = 4T
¨Ð¼Å¡Ò÷´Åͧ¾ºÇèÒ à¾Íà«ÍÃì äÇàÇÍÃì ä·ÁÁÔ § ÃԤѿàÇÍÃÕ ÁÕ »ÃÐÊÔ·¸ÔÀÒ¾´Õ ¡ÇèÒ ä·ÁÁÔ §
ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä» (à¹× ͧ¨Ò¡ ãªé
Eb /N0
¹éÍ¡ÇèÒ ã¹¡ÒÃ·Õ ¨Ð·ÓãËéÃкºÁÕ BER =
à·èҡѹ) â´Â੾ÒÐÍÂèÒ§ÂÔ § àÁ× Í·Ó§Ò¹·Õ ÃдѺ¤ÇÒÁÃعáç¢Í§ä·ÁÁÔ §¨ÔµàµÍÃì
σw /T
10−4
ÊÙ§
㹡ÒõÃǨÊͺÍѵÃÒ¡ÒÃÅÙèà¢éÒ ¢Í§ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áººµèÒ§æ ¨Ðãªé Ẻ¨ÓÅͧã¹ÃÙ» ·Õ 2.2 â´Â ¡Ó˹´ãËé
σw /T = 0%, τ̂0 = 0.5T ,
PLL ¨Ð¶Ù¡Í͡ẺÁÒÊÓËÃѺ
C
ÍͿ૵·Ò§¤ÇÒÁ¶Õ à·èҡѺ 0%, áÅФèÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã
= 50 ¹Ñ ¹¤×Í
α50
ÊÓËÃѺ
d
= 0 áÅÐ
4T
¤×Í 0.058 áÅÐ 0.049
µÒÁÅӴѺ ÃÙ»·Õ 2.14 à»ÃÕºà·ÕºÍѵÃÒ¡ÒÃÅÙèà¢éҢͧä·ÁÁÔ §ÃԤѿàÇÍÃÕẺµèÒ§æ â´Â¤Ô´à©ÅÕ Â¨Ò¡¡ÅØèÁ ¢éÍÁÙÅ 50000 ¡ÅØèÁ â´Â·Õ Ãкº¨Ð¶Ù¡¾Ô¨ÒóÒÇèÒ»ÃÐʺ¤ÇÒÁÊÓàÃç¨ã¹¡ÒÃÅÙèà¢éÒ ³ àÇÅÒ·Õ
k
¡çµèÍàÁ× Í
40
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
100
Genie−aided detector
Percentage of convergence
90 80 70
PSP−MM
60
Conventional timing recovery with hard decision (d = 0)
50 40 30
Conventional timing recovery with tentative decision (d = 4)
20 10 50
100
150
200
250
300
350
400
450
500
Time (in bit periods)
ÃÙ»·Õ 2.14: ÍѵÃÒ¡ÒÃÅÙèà¢éҢͧä·ÁÁÔ §ÃԤѿàÇÍÃÕẺµèÒ§æ àÁ× Íãªé
τ̂i
ÊÓËÃѺ
i≥k
ÁÕ ¤èÒ à·èÒ ¡Ñº ¤èÒ 0 ËÃ×Í
T
α50
·Õ
Eb /N0 = 10
´éǤèÒ ¤ÇÒÁ¤ÅÒ´à¤Å× Í¹ÂÔ¹ÂÍÁ
±10%
ÇèÒ Ãкº·Õ ãªé Genie aided detector ¨ÐÅÙèà¢éÒÀÒÂã¹ 50 ºÔµ «Ö §¨ÐÊÍ´¤Åéͧ¡Ñº ǧ¨Ã PLL ¢Í§ Genie aided detector ãªé¤èÒ·Õ ¶Ù¡µéͧ (¹Ñ ¹¤×Í
r̂k
=
rk )
α50
dB
¨Ò¡ÃÙ» ¨Ð¾º ·Õ ãªé à¾ÃÒÐÇèÒ
㹡ÒûÃѺ¤èÒÍͿ૵
·Ò§à¿Ê¢Í§¡ÒêѡµÑÇÍÂèÒ§µÑÇ¶Ñ´ä» ¹Í¡¨Ò¡¹Õ Âѧ¾ºÇèÒ à¾Íà«ÍÃìäÇàÇÍÃìä·ÁÁÔ §ÃԤѿàÇÍÃÕÁÕÍѵÃÒ¡Òà ÅÙèà¢éÒ·Õ ÃÇ´àÃçÇ àÁ× Íà·Õº¡Ñºä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä» »ÃÐÊÔ·¸ÔÀÒ¾¢Í§à¾Íà«ÍÃì äÇàÇÍÃì ä·ÁÁÔ § ÃԤѿàÇÍÃÕ ÊÒÁÒö·Õ ¨Ð·ÓãËé à¾Ô Á ¢Ö ¹ ä´é ÍÕ¡ â´Â¡ÒùÓä» ãªé §Ò¹ÃèÇÁ¡Ñ¹ ¡Ñº ÃËÑÊ á¡é䢢éͼԴ¾ÅÒ´ (ECC: error correction code) «Ö § ¼ÅÅѾ¸ì ·Õ ä´é ¨ÐàÃÕ¡ÇèÒ à¾Íà«ÍÃì äÇàÇÍÃì ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº·Ó§Ò¹« Ó (per survivor iterative timing recovery) [30, 39, 40] â´Â¨Ðãªé §Ò¹¡Ñº Ãкº·Õ ¶Ù¡ à¢éÒ ÃËÑÊ (coded system) «Ö § ¨Ò¡¼Å¡Ò÷´Åͧ¾ºÇèÒ à¾Íà«ÍÃì äÇàÇÍÃì ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº·Ó§Ò¹« Ó ÁÕ »ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡¡ÇèÒ ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº·Õ ãªé ¡Ñ¹ ·Ñ Çä»ÁÒ¡ â´Â੾ÒÐÍÂèÒ§ÂÔ § àÁ× Í ÃкºÁÕ ¢éͼԴ¾ÅÒ´·Ò§àÇÅÒÁÒ¡ ËÃ×Í àÁ× Í ·Ó§Ò¹ã¹Ãкº·Õ µéͧ¡ÒÃÍѵÃÒ¡Òà ÅÙèà¢éÒ·Õ ÃÇ´àÃçÇ ÊÓËÃѺ¼Ùéʹã¨ÊÒÁÒöÈÖ¡ÉÒÃÒÂÅÐàÍÕ´à¾Ô ÁàµÔÁä´éã¹ [30, 39, 40] «Ö §ÊÒÁÒö´Òǹì âËÅ´àÍ¡ÊÒÃàËÅèÒ¹Õ ä´é·Õ http://home.npru.ac.th/∼t3058
2.7.
ÊÃØ»·éÒº·
2.7
41
ÊÃØ»·éÒº·
㹺·¹Õ ä´é͸ԺÒ¶֧ËÅÑ¡¡Ò÷ӧҹ¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä» ÃÇÁ件֧ÇÔ¸Õ¡ÒÃÍ͡Ẻ ¤èÒ ¾ÒÃÒÁÔàµÍÃì ¢Í§Ç§¨Ã PLL â´Âãªé Ẻ¨ÓÅͧǧ¨Ãà¿ÊÅçÍ¡ÅÙ» ẺàªÔ§ àÊé¹ ¨Ò¡¼Å¡Ò÷´Åͧ¾º ÇèÒ ¤èÒ ¾ÒÃÒÁÔàµÍÃì ¢Í§Ç§¨Ã PLL ·Õ ä´é ¨Ò¡¡ÒÃÍ͡ẺµÒÁࡳ±ì ·Õ ¡Ó˹´äÇé ÁÕ ¤èÒ à´ÕÂǡѹ äÁèÇèÒ ã¹ Ãкº¨ÐÁÕ »ÃÔÁÒ³ÍͿ૵·Ò§¤ÇÒÁ¶Õ à·èÒã´ áÅСè͹·Õ ¨Ð¹Ó¤èÒ ¾ÒÃÒÁÔàµÍÃì ¢Í§Ç§¨Ã PLL ·Õ ä´é ¨Ò¡ ÇÔ¸Õ¡ÒÃÍ͡ẺµÒÁ·Õ ͸ԺÒÂ㹺·¹Õ ä»ãªé §Ò¹ ¼Ùéãªé ¨Ðµéͧ·ÓãËé ¤ÇÒÁªÑ¹ ¢Í§àÊé¹â¤é§ ÃÙ» µÑÇ àÍʢͧ ǧ¨Ã TED ÁÕ¤èÒà» ¹¤èÒË¹Ö § ³ ¨Ø´¡Óà¹Ô´¡è͹àÊÁÍ ¹Í¡¨Ò¡¹Õ ¼Å¡Ò÷´ÅͧáÊ´§ãËéàËç¹ÇèÒ ä·ÁÁÔ § ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä»·Ó§Ò¹ä´éäÁè´Õ àÁ× ÍÃкºÁÕ¢éͼԴ¾ÅÒ´·Ò§àÇÅÒÁÒ¡ ËÃ×ÍàÁ× Í·Ó§Ò¹ã¹Ãкº ·Õ µéͧ¡ÒÃÍѵÃÒ¡ÒÃÅÙèà¢éÒ·Õ ÃÇ´àÃçÇ ÍÂèÒ§äáçµÒÁã¹·Ò§»¯ÔºÑµÔ ä·ÁÁÔ § ÃԤѿàÇÍÃÕ áºº·Õ ãªé ¡Ñ¹ ·Ñ Ç仡ç Âѧ à» ¹ ·Õ ¹ÔÂÁãªé §Ò¹¡Ñ¹ ÁÒ¡ã¹ à¡×ͺ¨Ð·Ø¡§Ò¹»ÃÐÂØ¡µì à¹× ͧ¨Ò¡à» ¹Ç§¨Ã·Õ §èÒµèÍ¡ÒÃÊÃéÒ§áÅÐÊÒÁÒö·Ó§Ò¹ä´é´Õà¾Õ§¾Í ¶éÒàÅ×Í¡ ãªé¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL ·Õ àËÁÒÐÊÁ ÃÇÁ·Ñ §ÃкºÁÕ¢éͼԴ¾ÅÒ´·Ò§àÇÅÒäÁèÁÒ¡¹Ñ¡ áÅÐÃкº äÁèÁÕ¤ÇÒÁµéͧ¡ÒÃÍѵÃÒ¡ÒÃÅÙèà¢éÒ·Õ ÃÇ´àÃçÇ
2.8
à຺½ ¡ËÑ´·éÒº·
1. ¨§Í¸ÔºÒÂËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨Ãà¿ÊÅçÍ¡ÅÙ» (PLL: phased lock loop) ÁÒ¾ÍÊѧࢻ
2. ¨§Í¸ÔºÒ墄 ¹µÍ¹¡ÒÃÍ͡Ẻ¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL Íѹ´Ñº·Õ Ë¹Ö §
3. ¨§Í¸ÔºÒ墄 ¹µÍ¹¡ÒÃÍ͡Ẻ¤èÒ¾ÒÃÒÁÔàµÍÃì¢Í§Ç§¨Ã PLL Íѹ´Ñº·Õ Êͧ
4. ¨§Í¸ÔºÒ¤سÊÁºÑµÔáÅлÃÐ⪹ì¢Í§àÊé¹â¤é§ÃÙ»µÑÇàÍÊ
5. ¨§¤Ó¹Ç³ËÒÊÁ¡ÒÃàÊé¹â¤é§ ÃÙ» µÑÇ àÍʢͧǧ¨Ã M&M TED áÅÐÇÒ´ÃÙ» ä·ÁÁÔ § ¿ §¡ìªÑ¹ ¢Í§ ªèͧÊÑÒ³
H(D)
µèÍ仹Õ
5.1)
H(D)
=
1−D
5.2)
H(D)
=
1+D
42
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
5.3)
H(D)
=
1 + 2D + D2
5.4)
H(D)
=
1 + D − D2 − D3
5.5)
H(D)
=
1 + 3D + 3D2 + D3
6. ¨§à»ÃÕºà·ÕºËÅÑ¡¡Ò÷ӧҹ¢Í§ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä»áÅÐä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ »ÃÐÁÒ³¤èÒ㹪èǧ ÁÒ¾ÍÊѧࢻ
º··Õ 3
¡ÒÃÍÍ¡à຺·ÒÃìà¡çµààÅÐÍÕ¤ÇÍäÅà«ÍÃì
㹺·¹Õ ¨Ð͸ԺÒ¶֧ ÇÔ¸Õ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ (target) áÅÐÍÕ¤ÇÍäÅà«ÍÃì (equalizer) ãËé àËÁÒÐÊÁ ¡Ñº ªèͧÊÑÒ³¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì â´Â·Õ ·ÒÃìà¡çµ·Õ ´Õ ¨ÐµéͧÁÕ ¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ·Õ ã¡Åéà¤Õ§ ¡Ñº ¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§ªèͧÊÑÒ³ãËé ÁÒ¡·Õ ÊØ´ ã¹·Ò§»¯ÔºÑµÔ ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ áÅÐ ÍÕ¤ÇÍäÅà«ÍÃì ÊÒÁÒö·Óä´é ËÅÒÂÇÔ¸Õ áµè 㹺·¹Õ ¨Ð¡ÅèÒǶ֧ ੾ÒÐÇÔ¸Õ¡Òà ¢éͼԴ¾ÅÒ´¡ÓÅѧ Êͧà©ÅÕ Â ·Õ ¹éÍÂÊØ´ (MMSE: minimum mean squared error) à·èÒ¹Ñ ¹ [19] à¹× ͧ¨Ò¡ à» ¹ÇÔ¸Õ·Õ §èÒµèÍ¡ÒÃ¹Ó ä»ãªé §Ò¹¨ÃÔ§ ¾ÃéÍÁ·Ñ § à»ÃÕºà·Õº¼Å¡Ò÷´Åͧ·Õ ä´é ¨Ò¡¡ÒÃãªé ·ÒÃìà¡çµ ẺµèÒ§æ 㹪èͧÊÑÒ³ ¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì
3.1
º·¹Ó
ã¹Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ (Viterbi detector) [15] ¨Ð¶Ù¡ ¹ÓÁÒãªé §Ò¹ÃèÇÁ¡Ñ¹ ¡Ñº ÍÕ¤ÇÍäÅà«ÍÃì Ẻ PR
(partial response equalizer) «Ö § ÍÕ¤ÇÍäÅà«ÍÃì
¹Õ ¡ç¤×Í Ç§¨Ã¡ÃͧẺàªÔ§àÊé¹ (linear lter) ·Õ ·Ó˹éÒ·Õ ã¹¡ÒûÃѺÃÙ»ÃèÒ§¢Í§¼ÅµÍºÊ¹Í§ÃÇÁ¢Í§ 1
ªèͧÊÑÒ³ãËéÍÂÙèã¹ÃÙ»¢Í§¼ÅµÍºÊ¹Í§·Õ µéͧ¡Òà ËÃ×Í·Õ àÃÕ¡¡Ñ¹ÇèÒ ·ÒÃìà¡çµ 1
(target) ¨Ò¡¹Ñ ¹
·ÒÃìà¡çµ ¤×Í Ç§¨Ã¡ÃͧẺàªÔ§ àÊé¹ ·Õ ÁÕ ¨Ó¹Ç¹á·ç» (tap) ¹éÍ àÁ× Í à·Õº¡Ñº ¨Ó¹Ç¹á·ç» ¢Í§ªèͧÊÑÒ³ áµè ÁÕ ¼Å
µÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ ã¡Åéà¤Õ§¡Ñº¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ ¢Í§ªèͧÊÑÒ³
43
44
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¡ç ¨Ð¹ÓàÍÒ¢éÍÁÙÅ àÍÒµì¾Øµ ¢Í§ÍÕ¤ÇÍäÅà«ÍÃì ÁÒ·Ó¡ÒõÃǨËÒÅӴѺ (sequence detection) Ẻ¤ÇèÐà» ¹ ÁÒ¡ÊØ´ (ML: maximum likelihood) à¾× Í ¤Ó¹Ç³ËÒÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ ·Õ Êè§ ÁҨҡǧ¨ÃÀÒ¤Êè§ ¢Ñ ¹µÍ¹·Ñ § 2 ¹Õ ÃÇÁàÃÕ¡¡Ñ¹ ÇèÒ à·¤¹Ô¤ ¼ÅµÍºÊ¹Í§ºÒ§Êèǹ¤ÇèÐà» ¹ ÁÒ¡ÊØ´ (PRML: partial response maximum likelihood) «Ö § ¶×Í ÇèÒ à» ¹ ËÑÇã¨ÊÓ¤Ñ ¢Í§Ãкº¡Òà »ÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ìã¹» ¨¨ØºÑ¹ ·ÒÃìà¡çµáºº PR ·Õ à» ¹·Õ ÂÍÁÃѺ¡Ñ¹ã¹Ãкº¡Òúѹ·Ö¡áººá¹Ç¹Í¹ (longitudinal recording) ¨ÐÍÂÙèã¹ÃÙ»¢Í§¾ËعÒÁ (polynomial)
H(D) = (1 − D)(1 + D)n â´Â·Õ
n
¤×Í àÅ¢¨Ó¹Ç¹àµçÁºÇ¡ áÅÐ
D
(3.1)
¤×Í µÑÇ´Óà¹Ô¹¡ÒÃ˹èǧàÇÅÒ (delay operator) ¨Ò¡ÊÁ¡ÒÃ
(3.1) ¨Ð¾ºÇèÒ ¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ ¢Í§·ÒÃìà¡çµ¹Õ ¨ÐÁÕÊ໡µÃÑÁ¤èÒÈÙ¹Âì (spectral null) ³ ¤ÇÒÁ¶Õ ¤èÒÈÙ¹ÂìáÅФÇÒÁ¶Õ 乤ÇÔµÊì (Nyquist frequency) à¹× ͧ¨Ò¡ ÁÕ¾¨¹ì·Õ à» ¹
(1−D)
áÅÐ
(1+D)
µÒÁ
ÅӴѺ ã¹¢³Ð·Õ ÊÑÒ³ read back ¢Í§Ãкº¡Òúѹ·Ö¡áººá¹ÇµÑ § (perpendicular recording) ¨ÐÁÕͧ¤ì»ÃСͺ¢Í§ä¿¿ Ò¡ÃÐáʵç (d.c. component) ´Ñ§¹Ñ ¹ ¾¨¹ì
(1 − D)
¨Ö§äÁè¨Óà» ¹ÊÓËÃѺ
Ãкº¡Òúѹ·Ö¡ Ẻá¹ÇµÑ § à¾ÃÒÐ©Ð¹Ñ ¹ ·ÒÃìà¡çµ Ẻ PR ·Õ à» ¹ ·Õ ÂÍÁÃѺ ã¹Ãкº¡Òúѹ·Ö¡ Ẻ á¹ÇµÑ § ¨ÐÍÂÙèã¹ÃÙ»¢Í§¾ËعÒÁ
H(D) = (1 + D)n
(3.2)
µÒÃÒ§·Õ 3.1 áÊ´§·ÒÃìà¡çµ Ẻ PR ·Õ à» ¹ ·Õ ÂÍÁÃѺ ¡Ñ¹ ·Ñ Çä» ã¹Ãкº¡Òúѹ·Ö¡ Ẻá¹Ç¹Í¹ áÅÐẺá¹ÇµÑ § áÅÐÃÙ»·Õ 3.1 áÊ´§¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ ¢Í§·ÒÃìà¡çµáººµèÒ§æ àÁ× Íà»ÃÕºà·Õº ¡Ñº ¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§ªèͧÊÑÒ³·Õ ND = 2 áÅÐ 2.5 ¨Ò¡ÃÙ» ·Õ 3.1 ¨Ð¾ºÇèÒ àÁ× Í ¤èÒ ND ¢Í§ªèͧÊÑÒ³ÁÕ ¤èÒ à¾Ô Á ¢Ö ¹ ·ÒÃìà¡çµ·Õ ãªé ¡ç ¤ÇÃ·Õ ¨ÐÁÕ ¨Ó¹Ç¹á·ç» ÁÒ¡¢Ö ¹ (¤èÒ
n
ÁÒ¡¢Ö ¹)
à¾× Í·Õ ¨Ð·ÓãËé ÁÕ ¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ·Õ ÊÍ´¤Åéͧ¡Ñº ¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§ªèͧÊÑÒ³ãËé ÁÒ¡·Õ ÊØ´ ÍÂèÒ§äáçµÒÁ ·ÒÃìà¡çµ·Õ ãªé äÁè ¤ÇÃ·Õ ¨ÐÁÕ ¨Ó¹Ç¹á·ç» ÁÒ¡à¡Ô¹ ¤ÇÒÁ¨Óà» ¹ à¾ÃÒШÐÊè§ ¼Å·Ó ãËéǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔÁÕ¤ÇÒÁ«Ñº«é͹ (complexity) ÁÒ¡¢Ö ¹ «Ö §¨Ð͸ԺÒµèÍä»ã¹º··Õ 4 ¨Ò¡ÊÁ¡Òà (3.1) áÅÐ (3.2) Êѧࡵ¨Ð¾ºÇèÒ·ÒÃìà¡çµáºº PR ¨ÐÁÕ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§áµèÅÐá·ç» à» ¹àÅ¢¨Ó¹Ç¹àµçÁ áµè¶éÒÃкº PRML ÂÍÁãªé·ÒÃìà¡çµ·Õ ÁÕ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§áµèÅÐá·ç»à» ¹àÅ¢¨Ó¹Ç¹ ¨ÃÔ§¨Ð¾ºÇèÒ·ÒÃìà¡çµáºº¹Õ ¨ÐÊÒÁÒöªèÇÂà¾Ô Á »ÃÐÊÔ·¸ÔÀÒ¾ÃÇÁ¢Í§Ãкºä´é ÁÒ¡ â´Â੾ÒÐÍÂèÒ§ÂÔ §
3.1.
º·¹Ó
45
µÒÃÒ§·Õ 3.1: µÑÇÍÂèÒ§·ÒÃìà¡çµáºº PR ·Õ à» ¹·Õ ÂÍÁÃѺ¡Ñ¹ã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì
n=1
·ÒÃìà¡çµáºº PR Ãкº¡Òúѹ·Ö¡áººá¹Ç¹Í¹
PR4
n=2
[1 0 − 1]
1 − D2 Ãкº¡Òúѹ·Ö¡áººá¹ÇµÑ §
PR1
EPR4
n=3
[1 1 − 1 − 1]
1 + D − D2 − D3
[1 1]
PR2
[1 2 1]
[1 2 0 − 2 − 1]
1 + 2D − 2D3 − D4 EPR2
1 + 2D + D2
1+D
EEPR4
[1 3 3 1]
1 + 3D + 3D2 + D3
·Õ ND ÊÙ§æ â´Â·ÒÃìà¡çµ ÅÑ¡É³Ð¹Õ ¨ÐàÃÕ¡¡Ñ¹ ·Ñ Çä»ÇèÒ ·ÒÃìà¡çµáºº GPR (generalized partial response) «Ö §ÊÒÁÒö·Õ ¨ÐËÒä´é¨Ò¡ËÅÒÂÇÔ¸Õ¡Òà àªè¹ [41, 42, 43, 44, 45, 46]
1) ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ â´ÂãËé ¼ÅµÍºÊ¹Í§·ÒÃìà¡çµ (target response) ÁÕ ÃÙ»ÃèÒ§àËÁ×͹¡Ñº ¼Å µÍºÊ¹Í§¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð (transition response) ËÃ×ͼŵͺʹͧ䴺Ե (dibit response) ¢Í§ªèͧÊÑÒ³ ·Ñ §ã¹â´àÁ¹àÇÅÒ (time domain) áÅÐâ´àÁ¹¤ÇÒÁ¶Õ (frequency domain) ÇÔ¸Õ¡ÒÃ¹Õ ¨Ð·Ó¡ÒÃËÒ·ÒÃìà¡çµ·Õ ÁÕÃÙ»ÃèÒ§¢Í§¼ÅµÍºÊ¹Í§·ÒÃìà¡çµàËÁ×͹¡Ñº¼ÅµÍºÊ¹Í§ ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹ÐËÃ×Í ¼ÅµÍºÊ¹Í§ä´ºÔµ¢Í§ªèͧÊÑÒ³áµèÅÐ ND ·Ñ § ã¹â´àÁ¹àÇÅÒáÅÐ â´àÁ¹¤ÇÒÁ¶Õ µÑÇÍÂèÒ§àªè¹ã¹Ãкº¡Òúѹ·Ö¡áººá¹ÇµÑ § ·ÒÃìà¡çµ·Õ ÊÍ´¤Åéͧ¡Ñº¼ÅµÍºÊ¹Í§ ä´ºÔµ [43] ã¹â´àÁ¹àÇÅÒ ä´éá¡è [1
−4
1], [1
−2 −2
1], áÅÐ [1 0
−8
0 1] à» ¹µé¹
¹Í¡¨Ò¡¹Õ ·ÒÃìà¡çµ ºÒ§áººÍÒ¨¨ÐäÁè ÊÍ´¤Åéͧ¡Ñº ¼ÅµÍºÊ¹Í§¢Í§ªèͧÊÑÒ³ã¹â´àÁ¹ àÇÅÒ áµè¨ÐÊÍ´¤Åéͧ¡Ñº¼ÅµÍºÊ¹Í§¢Í§ªèͧÊÑÒ³ã¹â´àÁ¹¤ÇÒÁ¶Õ ÁÒ¡¡çä´é ã¹·Ò§»¯ÔºÑµÔ ¤ÇÒÁÊÍ´¤Åéͧ¡Ñº ¼ÅµÍºÊ¹Í§¢Í§ªèͧÊÑÒ³ã¹â´àÁ¹¤ÇÒÁ¶Õ à» ¹ ÊÔ § ·Õ µéͧ¡ÒÃÁÒ¡¡ÇèÒ ¤ÇÒÁÊÍ´¤Åéͧã¹â´àÁ¹àÇÅÒ à¹× ͧ¨Ò¡ ¨ÐªèǺ͡ãËé ·ÃÒº¶Ö§ ¤Ø³ÊÁºÑµÔ à¡Õ ÂǡѺ ÍѵÃÒ¡Òà ¢ÂÒÂÊÑҳú¡Ç¹ (noise enhancement) ä´é
2) ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ ·Õ ·ÓãËé ¡ÓÅѧ ÃÇÁ¢Í§ÊÑҳú¡Ç¹·Õ ´éÒ¹¢ÒÍÍ¡¢Í§ÍÕ¤ÇÍäÅà«ÍÃì ÁÕ ¤èÒ ¹éÍÂ·Õ ÊØ´ ÇÔ¸Õ¡ÒÃ¹Õ ¨ÐÊÁÁØµÔ ÇèÒ ¹Ñ¡Í͡ẺÃкº·ÃÒºÇèÒ ªèͧÊÑÒ³¤×Í ÍÐäà à¾× Í·Õ ¨Ðä´é ¹ÓàÍÒ
46
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
1.2
Channel (ND = 2) Channel (ND = 2.5) PR4 [1 0 −1] EPR4 [1 1 −1 −1] EEPR4 [1 2 0 −2 −1]
Normalized magnitude
1.0
0.8
0.6
0.4
0.2
0.0 0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
(a) Normalized frequency (fT) 1.0
Channel (ND = 2) Channel (ND = 2.5) PR2 [1 2 1] EPR2 [1 3 3 1] EEPR2 [1 4 6 4 1]
Normalized magnitude
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
(b) Normalized frequency (fT)
ÃÙ»·Õ 3.1: ¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ ¢Í§·ÒÃìà¡çµáººµèÒ§æ ÊÓËÃѺÃкººÑ¹·Ö¡ (a) Ẻá¹Ç¹Í¹ áÅÐ (b) Ẻá¹ÇµÑ §
ªèͧÊÑÒ³¹Ñ ¹ ÁÒãªé 㹡ÒäӹdzËÒ¿ §¡ìªÑ¹ ¶èÒÂâ͹¢Í§ÍÕ¤ÇÍäÅà«ÍÃì ¨Ò¡¹Ñ ¹ ¡ç ¨Ð·Ó¡Òà ¤Ó¹Ç³ËÒ¡ÓÅѧ ÃÇÁ¢Í§ÊÑҳú¡Ç¹·Õ ´éÒ¹¢ÒÍÍ¡¢Í§ÍÕ¤ÇÍäÅà«ÍÃì áÅÐàÁ× Í ä´é ¡ÓÅѧ ÃÇÁ
3.2.
¡ÒÃÍÍ¡à຺·ÒÃìà¡çµ´éÇÂÇÔ¸Õ¡Òà MMSE
47
¢Í§ÊÑҳú¡Ç¹·Õ ÍÂÙè ã¹ÃÙ» ¢Í§ÊÁ¡Ò÷ҧ¤³ÔµÈÒʵÃì áÅéÇ ¡ç ¨Ð·Ó¡ÒÃËÒ¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§·ÒÃìà¡çµ ·Õ ·ÓãËé ¡ÓÅѧ ÃÇÁ¢Í§ÊÑҳú¡Ç¹ÁÕ ¤èÒ ¹éÍÂ·Õ ÊØ´ â´Âãªé à·¤¹Ô¤ ¡ÒÃËÒ͹ؾѹ¸ì (di erentiation) ÊÓËÃѺÃÒÂÅÐàÍÕ´¢Í§ÇÔ¸Õ¡ÒÃ¹Õ ÊÒÁÒöÈÖ¡ÉÒà¾Ô ÁàµÔÁä´é¨Ò¡ [41]
3) ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ ·Õ ·ÓãËé ¤èÒ SNR »ÃÐÊÔ·¸Ô¼Å (e ective SNR) ·Õ ´éÒ¹¢ÒÍÍ¡¢Í§Ç§¨Ã µÃǨËÒÇÕà·ÍÃìºÔÁÕ¤èÒÁÒ¡·Õ ÊØ´ ÇÔ¸Õ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ ẺµèÒ§æ µÒÁ·Õ ¡ÅèÒÇÁÒ¢éÒ§µé¹ ¹Õ äÁè ä´é ÃѺ»ÃСѹ ÇèÒ »ÃÐÊÔ·¸Ô ÀÒ¾ÃÇÁ¢Í§Ãкºã¹ÃÙ» ¢Í§ÍѵÃÒ¢éͼԴ¾ÅÒ´ºÔµ (BER: bit error rate) ·Õ ´éÒ¹¢ÒÍÍ¡¢Í§ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨ÐÁÕ ¤èÒ ¹éÍÂ·Õ ÊØ´ ã¹ [19, 42] ä´é àʹÍÇÔ¸Õ¡ÒÃËÒ ·ÒÃìà¡çµ·Õ àËÁÒÐ·Õ ÊØ´ (optimal target) ·Õ ¨Ð·ÓãËé ¤èÒ SNR »ÃÐÊÔ·¸Ô¼Å·Õ ´éÒ¹¢ÒÍÍ¡¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÁÕ ¤èÒ ÁÒ¡·Õ ÊØ´ ËÃ×ÍÍÕ¡¹ÑÂË¹Ö §¡ç¤×Í ·ÒÃìà¡çµ·Õ ·ÓãËé BER ¢Í§Ãкº àÁ× ÍÇÑ´·Õ ´éÒ¹¢ÒÍÍ¡¢Í§Ç§¨Ã µÃǨËÒÇÕà·ÍÃìºÔÁÕ¤èÒ¹éÍÂ·Õ ÊØ´ ÊÓËÃѺ¼Ùéʹã¨ÊÒÁÒöÈÖ¡ÉÒÃÒÂÅÐàÍÕ´à¾Ô ÁàµÔÁä´éã¹ [19, 42]
4) ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ·Õ ·ÓãËé¢éͼԴ¾ÅÒ´¡ÓÅѧÊͧà©ÅÕ Â (MSE: mean squared error) ÃÐËÇèÒ§ ÊÑÒ³·Õ ´éÒ¹¢ÒÍÍ¡¢Í§ÍÕ¤ÇÍäÅà«ÍÃìáÅÐÊÑÒ³·Õ µéͧ¡Òà (¹Ñ ¹¤×Í ÊÑÒ³µÒÁ·ÒÃìà¡çµ ·Õ µéͧ¡ÒÃ) ÁÕ¤èÒ¹éÍÂ·Õ ÊØ´ ¨Ò¡¡ÒÃÈÖ¡ÉÒ¾ºÇèÒ ÇÔ¸Õ¡ÒÃ¹Õ à» ¹ ÇÔ¸Õ¡ÒÃ·Õ §èÒÂáÅÐàËÁÒÐÊÓËÃѺ ¡ÒùÓÁÒãªé §Ò¹¨ÃÔ§ ã¹ ·Ò§»¯ÔºÑµÔ [19] ¹Í¡¨Ò¡¹Õ ·ÒÃìà¡çµ·Õ ä´é ¨ÐÁÕ »ÃÐÊÔ·¸ÔÀÒ¾ã¡Åéà¤Õ§¡Ñº ·ÒÃìà¡çµ ·Õ àËÁÒÐ·Õ ÊØ´ ÇÔ¸Õ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ Ẻ¹Õ ¨ÐÃÙé¨Ñ¡ ¡Ñ¹ ã¹ª× Í ÇèÒ ÇÔ¸Õ¡Òà ¢éͼԴ¾ÅÒ´¡ÓÅѧ Êͧà©ÅÕ Â ·Õ ¹éÍ ÊØ´ (MMSE: minimum mean squared error) «Ö §¨Ð͸ԺÒÂÃÒÂÅÐàÍÕ´ÇÔ¸Õ¡ÒÃ¹Õ ´Ñ§µèÍ仹Õ
3.2
¡ÒÃÍÍ¡à຺·ÒÃìà¡çµ´éÇÂÇÔ¸Õ¡Òà MMSE
¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ´éÇÂÇÔ¸Õ¡Òà MMSE [19] ¨Ð·ÓãËéäé´é·ÒÃìà¡çµËÅÒÂÃٻẺµÒÁà§× ͹䢺ѧ¤Ñº (con straint) ·Õ ¡Ó˹´Å§ä»ã¹ÃÐËÇèÒ§¡Ãкǹ¡ÒÃÍ͡Ẻ ãËé¾Ô¨ÒóÒẺ¨ÓÅͧÃкºã¹ÃÙ»·Õ 3.2 â´Â ·Õ ÍÕ¤ÇÍäÅà«ÍÃì¨Ð¾ÂÒÂÒÁÊÃéÒ§¢éÍÁÙÅàÍÒµì¾Øµ
yk
ãËéÁÕ¤èÒã¡Åéà¤Õ§¡Ñº¢éÍÁÙÅ·Õ µéͧ¡ÒÃ
rk
â´Â»ÃÒȨҡ¡ÒâÂÒÂÊÑҳú¡Ç¹ ¶éÒ¡Ó˹´ãËéÍÕ¤ÇÍäÅà«ÍÃìÁըӹǹá·ç»à·èҡѺ
ãËéÁÒ¡·Õ ÊØ´
N = 2K +1
48
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
n(t) equalizer
ak
1 − D bk g(t) {±1} 2
p(t)
∆t k
sk
s(t)
LPF
F(D)
t k = kT
y k Viterbi âk detector
target
rk
H(D)
wk
ÃÙ»·Õ 3.2: Ẻ¨ÓÅͧÊÓËÃѺ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ´éÇÂÇÔ¸Õ¡Òà MMSE
á·ç» áÅÐÊÁÁØµÔ ãËé á·ç» ÈÙ¹Âì¡ÅÒ§ÍÂÙè ·Õ àÇÅÒ ÍÂÙèã¹ÃÙ»ÊÁ¡Ò÷ҧ¤³ÔµÈÒʵÃìã¹â´àÁ¹
D
k =0
à¾ÃÒÐ©Ð¹Ñ ¹ ÍÕ¤ÇÍäÅà«ÍÃì ÊÒÁÒö·Õ ¨Ðà¢Õ¹ãËé
ä´é ¤×Í
K X
F (D) =
fk Dk
(3.3)
k=−K àÁ× Í
D
¤×Í µÑÇ´Óà¹Ô¹¡ÒÃ˹èǧàÇÅÒ
T
˹èÇ 㹷ӹͧà´ÕÂǡѹ ·ÒÃìà¡çµ·Õ Áըӹǹá·ç»à·èҡѺ
á·ç» ¡ç¨ÐÊÒÁÒöà¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§¿ §¡ìªÑ¹ã¹â´àÁ¹
H(D) =
L−1 X
D
L
ä´é ¤×Í
hk Dk
(3.4)
k=0 â´Â·Õ
fk
áÅÐ
hk
à» ¹¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ·Õ à» ¹àÅ¢¨Ó¹Ç¹¨ÃÔ§ã¹áµèÅÐá·ç»¢Í§ÍÕ¤ÇÍäÅà«ÍÃìáÅзÒÃìà¡çµ
µÒÁÅӴѺ ¨Ø´»ÃÐʧ¤ì 㹡ÒÃÍ͡Ẻ·ÒÃìà¡çµ ´éÇÂÇÔ¸Õ¡Òà MMSE ¤×Í ¨Ð·Ó¡ÒäӹdzËÒ¤èÒ ÊÑÁ»ÃÐ ÊÔ·¸Ô ¢Í§
F (D)
áÅÐ
H(D) yk
àÍÒµì¾Øµ ¢Í§ÍÕ¤ÇÍäÅà«ÍÃì ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô
fk
and
hk
ä» ¾ÃéÍÁ¡Ñ¹ ã¹ àÇÅÒ à´ÕÂǡѹ â´Â ¡Òà ·Ó ãËé ¤èÒ MSE ÃÐËÇèÒ§ ¢éÍÁÙÅ
áÅТéÍÁÙÅ àÍÒµì¾Øµ ¢Í§·ÒÃìà¡çµ
rk
ÁÕ ¤èÒ ¹éÍÂ·Õ ÊØ´ ËÃ×Í ÍÕ¡ ¹ÑÂ Ë¹Ö § ¤×Í
¨Ð¶Ù¡àÅ×Í¡ à¾× Í·Õ ·ÓãËé¤èÒ
£ ¤ £ ¤ E wk2 = E {(sk ∗ fk ) − (ak ∗ hk )}2 ÁÕ¤èÒ¹éÍÂ·Õ ÊØ´ àÁ× Í
wk
=
yk − rk
¤×Í ¢éͼԴ¾ÅÒ´·Õ ä´é¨Ò¡¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ,
¤Í¹âÇÅ٪ѹ (convolution operator), áÅÐ tor)
E[·]
(3.5)
∗
¤×Í µÑÇ´Óà¹Ô¹¡ÒÃ
¤×Í µÑÇ´Óà¹Ô¹¡ÒäèÒ¤Ò´ËÁÒ (expectation opera
3.2.
¡ÒÃÍÍ¡à຺·ÒÃìà¡çµ´éÇÂÇÔ¸Õ¡Òà MMSE
¶éÒ¡Ó˹´ãËéàÇ¡àµÍÃìá¹ÇµÑ § â´Â·Õ
hk
and
fk
H
49
[h0 , h1 , · · · , hL−1 ]T
=
¤×Í ¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§
H(D)
áÅÐ
F (D)
and
¢éÍÁÙÅ
sk , A
R
i
¤×Í àÁ·ÃÔ¡«ìÍѵÊËÊÑÁ¾Ñ¹¸ì¢¹Ò´
áÅÐá¹ÇµÑ §·Õ
j)
K = 10
[·]T
¤×Í à¤Ã× Í§ËÁÒÂ
(ÍÕ¤ÇÍäÅà«ÍÃìÁÕ·Ñ §ËÁ´ 21
¤×Í àÁ·ÃÔ¡«ìÍѵÊËÊÑÁ¾Ñ¹¸ì (auto correlation matrix) ¢¹Ò´
¢éÒÁ (cross correlation matrix) ¢¹Ò´ ·Õ
[f−K , · · · , f0 , · · · , fK ]T
=
µÒÁÅӴѺ, áÅÐ
àÁ·ÃÔ¡«ìÊÅѺà»ÅÕ Â¹ (transpose matrix) 㹺·¹Õ ¨Ð¡Ó˹´ãËé á·ç») ¶éÒ¡Ó˹´ãËé
F
L×L
N ×L
¢Í§¢éÍÁÙÅ
ÃÐËÇèÒ§¢éÍÁÙÅ
ak ,
sk
áÅÐ
áÅÐ
ak
P
N ×N
¢Í§
¤×Í àÁ·ÃÔ¡«ìÊËÊÑÁ¾Ñ¹¸ì
â´Â·Õ ÊÁÒªÔ¡
(i, j)
(á¶Ç
¢Í§àÁ·ÃÔ¡«ì·Ñ §ÊÒÁ¹Õ ¤×Í
R(i, j) = E
"S−1 X
# sk+K−i sk+K−j ,
k=0
A(i, j) = E P(i, j) = E
"S−1 X k=0 "S−1 X
−K ≤ i, j ≤ K
(3.6)
# ak−i ak−j ,
0 ≤ i, j ≤ L − 1
(3.7)
# sk+K−i ak−j ,
−K ≤ i ≤ K, 0 ≤ j ≤ L − 1
(3.8)
k=0
S
àÁ× Í
¤×Í ¤ÇÒÁÂÒÇ (ËÃ×ͨӹǹºÔµ) ¢Í§ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
{ak }
¨Ò¡µÑÇá»Ã·Õ ¡Ó˹´ãËé¢éÒ§µé¹¹Õ ÊÁ¡Òà (3.5) ÊÒÁÒö·Õ ¨Ðà¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§àÁ·ÃÔ¡«ìä´é ´Ñ§¹Õ
E[w2 ] = FT RF + HT AH − 2FT PH 㹡Ò÷ÓãËé¤èÒ
E[w2 ]
ÁÕ¤èÒ¹éÍÂ·Õ ÊØ´â´Âà·Õº¡Ñº
F
áÅÐ
H
(3.9)
¨ÐµéͧÁÕ¡ÒáÓ˹´à§× ͹䢺ѧ¤Ñºà¢éÒä»
ã¹ÃÐËÇèÒ§¡Ãкǹ¡Ò÷ÓãËé ÁÕ ¤èÒ ¹éÍÂÊØ´ (minimization process) à¾× Í·Õ ¨ÐËÅÕ¡àÅÕ Â§¡ÒÃä´é ¼ÅÅѾ¸ì à» ¹
F=0
áÅÐ
G
3.2.1
áÅÐ
H=0
ã¹Êèǹ¹Õ ¨Ð͸ԺÒ¶֧à§× ͹䢺ѧ¤Ñº·Õ ¹èÒʹ㨠·Õ ãªé㹡ÒäӹdzËÒ¤èÒ
F
´Ñ§µèÍ仹Õ
à§× ͹䢺ѧ¤Ñºà຺âÁ¹Ô¡ (h0
= 1)
à§× ͹䢺ѧ¤ÑºáººâÁ¹Ô¡ (monic constraint) ¨Ð¡Ó˹´ãËé¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§á·ç»µÑÇáá¢Í§·ÒÃìà¡çµ ÁÕ¤èÒà·èҡѺ¤èÒË¹Ö § ¹Ñ ¹¤×Í
h0 = 1
[19] ¶éÒ¡Ó˹´ãËéàÇ¡àµÍÃìá¹ÇµÑ §
I
ááÁÕ¤èÒà·èҡѺ¤èÒË¹Ö § ÊèǹÊÁÒªÔ¡µÑÇÍ× ¹æ ÁÕ¤èÒà·èҡѺ¤èÒÈÙ¹Âì ¡ÅèÒǤ×Í à§× ͹䢺ѧ¤ÑºáººâÁ¹Ô¡¹Õ ÊÒÁÒöà¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§àÁ·ÃÔ¡«ìä´é ¤×Í
¢¹Ò´
I
=
L×1
·Õ ÁÕÊÁÒªÔ¡µÑÇ
[1, 0, · · · , 0]T
IT H = 1
´Ñ§¹Ñ ¹
50
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¡Ãкǹ¡ÒÃ㹡ÒÃÍ͡Ẻ·ÒÃìà¡çµ ´éÇÂÇÔ¸Õ¡ÒÃ¹Õ ¤×Í ¡Ò÷ÓãËé ¤èÒ MSE ã¹ÊÁ¡Òà (3.9) ÁÕ ¤èÒ ¹éÍÂ·Õ ÊØ´ â´Â¾ÂÒÂÒÁÃÑ¡ÉÒãËé¤èÒ
IT H
=
1
ÍÂÙèµÅÍ´àÇÅÒ ¹Ñ ¹¤×Í ¡Ãкǹ¡ÒÃ¹Õ ¨Ð·ÓãËé¤èÒ
E[w2 ] = FT RF + HT AH − 2FT PH − 2λ(IT H − 1) ÁÕ ¤èÒ ¹éÍÂ·Õ ÊØ´ â´Â·Õ
λ
(3.10)
¤×Í µÑǤٳ ÅÒ¡ÃÒ¹¨ì (Lagrange multiplier) à» ¹¤èÒÊà¡ÅÒÃì (scalar) ¡ÒÃ
·ÓãËé ÊÁ¡Òà (3.10) ÁÕ ¤èÒ ¹éÍÂ·Õ ÊØ´ ÊÒÁÒö·Óä´é â´Â¡ÒÃËÒ͹ؾѹ¸ì (di erentiation) ¢Í§ÊÁ¡Òà (3.10) à·Õº¡Ñº F, H, áÅÐ
λ
µÒÁÅӴѺ «Ö §¨Ðä´é¼ÅÅѾ¸ì´Ñ§¹Õ (ÃÒÂÅÐàÍÕ´ã¹ÀÒ¤¼¹Ç¡ ¤)
¡ ¢ ∂ E[w2 ] ¡ ∂F ¢ ∂ E[w2 ] ¡ ∂H ¢ ∂ E[w2 ] ∂λ
= 2RF − 2PH
(3.11)
= 2AH − 2PT F − 2λI
(3.12)
= −2IT H + 2
(3.13)
¨Ò¡¹Ñ ¹¡ç¡Ó˹´ãËé ¼ÅÅѾ¸ì ¢Í§Í¹Ø¾Ñ¹¸ì ·Ñ §ËÁ´·Õ ä´é ¨Ò¡ÊÁ¡Òà (3.11) (3.13) ÁÕ ¤èÒ à» ¹ ¤èÒ ÈÙ¹Âì ¹Ñ ¹¤×Í ¡Ó˹´ãËéÊÁ¡Òà (3.11) ÁÕ¤èÒà·èҡѺ¤èÒÈÙ¹Âì ¨Ðä´éÇèÒ
2RF − 2PH = 0 RF = PH F = R−1 PH
(3.14)
áÅСÓ˹´ãËéÊÁ¡Òà (3.12) ÁÕ¤èÒà·èҡѺ¤èÒÈÙ¹Âì ¨Ðä´éÇèÒ
2AH − 2PT F − 2λI = 0 AH − PT F = λI á·¹¤èÒ
F
(3.15)
¨Ò¡ÊÁ¡Òà (3.14) ŧã¹ÊÁ¡Òà (3.15) ¨Ðä´é
¡ ¢ AH − PT R−1 PH = λI ¡ ¢ A − PT R−1 P H = λI ¡ ¢−1 I H = λ A − PT R−1 P
(3.16)
3.2.
¡ÒÃÍÍ¡à຺·ÒÃìà¡çµ´éÇÂÇÔ¸Õ¡Òà MMSE
51
㹷ӹͧà´ÕÂǡѹ ¡Ó˹´ãËéÊÁ¡Òà (3.13) ÁÕ¤èÒà·èҡѺ¤èÒÈÙ¹Âì ¨Ðä´éÇèÒ
−2IT H + 2 = 0 IT H = 1 á·¹¤èÒ
H
(3.17)
¨Ò¡ÊÁ¡Òà (3.16) ŧã¹ÊÁ¡Òà (3.17) ¨Ðä´é
¡ ¢−1 IT λ A − PT R−1 P I = 1 λ =
IT (A
1 − P R−1 P)−1 I T
(3.18)
à¾ÃÒÐ©Ð¹Ñ ¹â´ÂÊÃØ»áÅéÇ ¢Ñ ¹µÍ¹ã¹¡ÒÃÍ͡Ẻ·ÒÃìà¡çµáÅÐÍÕ¤ÇÍäÅà«ÍÃìÊÓËÃѺà§× ͹䢺ѧ¤Ñºà຺ âÁ¹Ô¡ (h0
= 1)
¤×Í
1) ¡Ó˹´¨Ó¹Ç¹á·ç»¢Í§·ÒÃìà¡çµáÅÐÍÕ¤ÇÍäÅà«ÍÃì (¤èÒ
L
áÅÐ
N)
¹Ñ ¹¤×ÍÊÃéèÒ§àÇ¡àµÍÃì F áÅÐ
H
2) ËÒ¤èÒàÁ·ÃÔ¡«ì R, A, áÅÐ P ¨Ò¡ÊÁ¡Òà (3.6) (3.8)
3) ÊÃéÒ§àÇ¡àµÍÃì I =
[1, 0, · · · , 0]T
4) ËÒ¤èÒµÑǤٳÅÒ¡ÃÒ¹¨ì
λ
¢¹Ò´
¨Ò¡ÊÁ¡Òà (3.18) ¹Ñ ¹¤×Í
1 − P R−1 P)−1 I
(3.19)
H = λ(A − PT R−1 P)−1 I
(3.20)
λ=
5) ËÒ¤èÒ·ÒÃìà¡çµ
H
L×1
IT (A
T
¨Ò¡ÊÁ¡Òà (3.16) ¹Ñ ¹¤×Í
6) ËÒ¤èÒÍÕ¤ÇÍäÅà«ÍÃì
F
¨Ò¡ÊÁ¡Òà (3.14) ¹Ñ ¹¤×Í
F = R−1 PH
(3.21)
52
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¤èÒ
λ
·Õ ä´é¨Ò¡ÊÁ¡Òà (3.19) à» ¹¤èÒ MMSE ·Õ ä´éÀÒÂãµéà§× ͹䢺ѧ¤Ñº¹Õ «Ö §ÊÒÁÒö¾ÔÊÙ¨¹ìä´éâ´Â
¡ÒÃá·¹¤èÒ F and H ¨Ò¡ÊÁ¡Òà (3.20) (3.21) ŧã¹ÊÁ¡Òà (3.9) ¹Ñ ¹¤×Í
E[w2 ] = (R−1 PH)T R(R−1 PH) + HT AH − 2(R−1 PH)T PH = HT PT (R−1 )T PH + HT AH − 2HT PT (R−1 )T PH = HT (A − PT R−1 P)H © ª = λ2 [IT (A − PT R−1 P)−1 ]T (A − PT R−1 P)(A − PT R−1 P)−1 I © ª = λ2 IT (A − PT R−1 P)−1 I = λ
(3.22)
ÊÁ¡Òà (3.22) ËÒä´éâ´ÂÍÒÈÑÂËÅÑ¡¤ÇÒÁ¨ÃÔ§·Õ ÇèÒ àÁ·ÃÔ¡«ì
R
áÅÐ
A
à» ¹àÁ·ÃÔ¡«ìÍѵÊËÊÑÁ¾Ñ¹¸ìẺ 2
ÊÁÁҵà ¹Í¡¨Ò¡¹Õ àÁ·ÃÔ¡«ì R, A áÅÐ P Âѧ໠¹àÁ·ÃÔ¡«ìẺ Toeplitz =
R−1
3.2.2
áÅÐ
£ ¤T (A − PT R−1 P)−1
à§× ͹䢺ѧ¤Ñºà຺
à§× ͹䢺ѧ¤Ñºáºº ¹Ñ ¹¤×Í
h1 = 1
=
«Ö §¨Ð·ÓãËéä´éÇèÒ
(R−1 )T
(A − PT R−1 P)−1
h1 = 1
h1 = 1 ¨Ð¡Ó˹´ãËé¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§á·ç»µÑÇÊͧ¢Í§·ÒÃìà¡çµÁÕ¤èÒà·èҡѺ¤èÒË¹Ö §
[19] Êèǹ¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§á·ç» µÑÇ Í× ¹æ ¨Ðà» ¹ ¤èÒ ÍÐäáçä´é â´Âà§× ͹䢺ѧ¤Ñº ¹Õ ÁÕ
Çѵ¶Ø»ÃÐʧ¤ìà¾× Íãªé㹡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§·ÒÃìà¡çµáººµèÒ§æ ¶éÒ¡Ó˹´ãËéàÇ¡àµÍÃìá¹Ç
J
µÑ §
J
=
¢¹Ò´
L×1
·Õ ÊÁÒªÔ¡µÑÇ·Õ ÊͧÁÕ¤èÒà·èҡѺ¤èÒË¹Ö § ÊèǹÊÁÒªÔ¡Í× ¹æ ÁÕ¤èÒà» ¹¤èÒÈÙ¹Âì ¡ÅèÒǤ×Í
[0, 1, 0, · · · , 0]T
´Ñ§¹Ñ ¹ à§× ͹䢺ѧ¤Ñº Ẻ
h1 = 1
ÊÒÁÒöà¢Õ¹ãËé ÍÂÙè ã¹ÃÙ» ¢Í§àÁ·ÃÔ¡«ì ä´é
JT H = 1
¤×Í
¡Ãкǹ¡ÒÃ㹡ÒÃÍ͡Ẻ·ÒÃìà¡çµâ´Âà§× ͹䢺ѧ¤Ñº¹Õ ¨ÐàËÁ×͹¡Ñº¡ÒÃÍ͡Ẻ·ÒÃìà¡çµâ´Âãªé à§× ͹䢺ѧ¤Ñº ẺâÁ¹Ô¡ à¾Õ§áµè à»ÅÕ Â¹¾¨¹ì ÊØ´·éÒÂã¹ÊÁ¡Òà (3.10) ¨Ò¡
2λ(JT H − 1) à» ¹àÇ¡àµÍÃì 2
2λ(IT H − 1)
ä»à» ¹
´Ñ§¹Ñ ¹ ¼ÅÅѾ¸ì·Õ ä´é¡ç¨ÐàËÁ×͹¡ÑºÊÁ¡Òà (3.19) (3.21) à¾Õ§áµèà»ÅÕ Â¹àÇ¡àµÍÃì
J
àÁ·ÃÔ¡«ìẺ Toeplitz ¤×Í àÁ·ÃÔ¡«ì·Õ ÊÁÒªÔ¡ã¹á¹ÇàÊé¹·á§ÁØÁà» ¹¤èÒ¤§·Õ ¤èÒà´ÕÂǡѹ
I
3.2.
¡ÒÃÍÍ¡à຺·ÒÃìà¡çµ´éÇÂÇÔ¸Õ¡Òà MMSE
3.2.3
53
à§× ͹䢺ѧ¤Ñºà຺¾Åѧ§Ò¹Ë¹Ö §Ë¹èÇ (H
T
H = 1)
à§× ͹䢺ѧ¤Ñº Ẻ¾Åѧ§Ò¹Ë¹Ö § ˹èÇ (unit energy) ÊÒÁÒöà¢Õ¹ãËé ÍÂÙè ã¹ÃÙ» ¢Í§àÁ·ÃÔ¡«ì ä´é ¤×Í
HT H
=
1
«Ö §à» ¹¡ÒáÓ˹´ãËé¾Åѧ§Ò¹¢Í§·ÒÃìà¡çµÁÕ¤èÒà·èҡѺ¤èÒË¹Ö § [19, 46]
¡Ãкǹ¡ÒÃ㹡ÒÃÍ͡Ẻ·ÒÃìà¡çµ ´éÇÂÇÔ¸Õ¡ÒÃ¹Õ ¤×Í ¡Ò÷ÓãËé ¤èÒ MSE ã¹ÊÁ¡Òà (3.9) ÁÕ ¤èÒ ¹éÍÂ·Õ ÊØ´ â´Â·Õ ¾ÂÒÂÒÁÃÑ¡ÉÒãËé¤èÒ
HT H
=
1
µÅÍ´àÇÅÒ ¡ÅèÒǤ×Í ¡Ãкǹ¡ÒÃ¹Õ ¨Ð·ÓãËé¤èÒ
E[w2 ] = FT RF + HT AH − 2FT PH − 2λ(HT H − 1) ÁÕ¤èÒµ ÓÊØ´ «Ö §ÊÒÁÒö·Óä´éâ´Â¡ÒÃËÒ͹ؾѹ¸ì¢Í§ÊÁ¡Òà (3.23) à·Õº¡Ñº F, H, áÅÐ
(3.23)
λ
µÒÁÅӴѺ
áÅéÇ ¡Ó˹´ãËé ¼ÅÅѾ¸ì ¢Í§Í¹Ø¾Ñ¹¸ì ·Ñ §ËÁ´·Õ ä´é ÁÕ ¤èÒ à» ¹ ¤èÒ ÈÙ¹Âì â´ÂàÁ× Í ·ÓµÒÁ¢Ñ ¹µÍ¹¹Õ áÅéÇ ¨Ð¾º ÇèÒ ·ÒÃìà¡çµáÅÐÍÕ¤ÇÍäÅà«ÍÃì ÊÒÁÒöËÒä´é¨Ò¡¡ÒÃá¡éÊÁ¡ÒÃ
(A − PT R−1 P)H = λH
(3.24)
«Ö §¡ÒÃá¡éÊÁ¡ÒÃ¹Õ ÁÕÅѡɳФÅéÒ¡Ѻ ¡ÒÃá¡é» ËÒ¤èÒÅѡɳÐ੾ÒÐ (eigenvalue problem) [12, 25] â´Â·Õ ¤èÒ (3.24) ¤èÒ
λ
λ
ã¹ÊÁ¡Òà (3.24) ÊÒÁÒö¾ÔÊÙ¨¹ìä´éÇèÒ ¤×Í ¤èÒ MMSE ¹Ñ ¹àͧ ´Ñ§¹Ñ ¹ ¨Ò¡ÊÁ¡ÒÃ
¨ÃÔ§æ áÅéÇ ¡ç ¤×Í ¤èÒ ÅѡɳÐ੾ÒÐ·Õ ¹éÍÂ·Õ ÊØ´ (minimum eigenvalue) ¢Í§àÁ·ÃÔ¡«ì
(A − PT R−1 P), H
¤×Í àÇ¡àµÍÃì ÅѡɳÐ੾ÒÐẺ¹ÍÃì ÁÍÅäÅ«ì(normalized eigenvector) ·Õ
ÊÍ´¤Åéͧ¡Ñº¤èÒÅѡɳÐ੾ÒÐ·Õ ¹éÍÂ·Õ ÊØ´, áÅÐ
3.2.4
F
¨ÐÁÕ¤èÒàËÁ×͹¡ÑºÊÁ¡Òà (3.21)
à§× ͹䢺ѧ¤Ñºà຺·ÒÃìà¡çµà©¾ÒÐ
ÊÓËÃѺ à§× ͹䢺ѧ¤Ñº Ẻ·ÒÃìà¡çµ ੾ÒÐ¹Õ ( xed target) ·ÒÃìà¡çµ¨Ð¶Ù¡ ¡Ó˹´ÁÒãËé µÑ §áµè àÃÔ Ááá ÊÔ § ·Õ Ãкºµéͧ¡Òà ¤×Í ÍÕ¤ÇÍäÅà«ÍÃì
F
·Õ àËÁÒÐÊÁ¡Ñº ·ÒÃìà¡çµ ·Õ ¡Ó˹´ÁÒãËé «Ö § ÊÒÁÒöËÒä´é ¨Ò¡
ÊÁ¡Òà (3.21) ¹Ñ ¹¤×Í
F = R−1 PG
(3.25)
â´Â¨ÐÃѺ»ÃСѹä´éÇèÒ¤èÒ MMSE ·Õ ä´é¨ÐÁÕ¤èÒà·èҡѺ¤èÒã¹ÊÁ¡Òà (3.9) ÇÔ¸Õ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµáºº¹Õ ÁÕ»ÃÐ⪹ìÁÒ¡ à¹× ͧ¨Ò¡ã¹¡ÒÃãªé§Ò¹¨ÃÔ§ ªÔ»»ÃÐÁÇżÅÊÑÒ³ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ¨Ð¶Ù¡ Í͡ẺÁÒãËé ãªé §Ò¹¡Ñº ·ÒÃìà¡çµ Ẻã´áººË¹Ö § à·èÒ¹Ñ ¹ ´Ñ§¹Ñ ¹ ¼Ùéãªé §Ò¹
54
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¨Ðµéͧ¾ÂÒÂÒÁ»ÃѺ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô áµèÅÐá·ç»¢Í§ÍÕ¤ÇÍäÅà«ÍÃìãËéÊÍ´¤Åéͧ¡Ñº·ÒÃìà¡çµ·Õ ¡Ó˹´ÁÒãËé à¾× ÍãËé ä´é ÊÑÒ³·Õ ´éÒ¹¢ÒÍÍ¡¢Í§ÍÕ¤ÇÍäÅà«ÍÃìé ÁÕ ÃÙ»ÃèÒ§àËÁ×͹¡Ñº ÊÑÒ³·Õ µéͧ¡Òà «Ö § ¨ÐªèÇÂ·Ó ãËé ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ·Ó§Ò¹ä´é §èÒ墅 ¹ ã¹·Ò§µÃ§¡Ñ¹¢éÒÁ ¶éÒ ·Ó¡ÒûÃѺ ¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô áµèÅÐá·ç» ¢Í§ÍÕ¤ÇÍäÅà«ÍÃìẺÅͧ¼Ô´Åͧ¶Ù¡ (trial and error) ¡ç¨Ð·ÓãËéàÊÕÂàÇÅÒÁÒ¡ áÅÐÍÕ¤ÇÍäÅà«ÍÃì·Õ ä´é ¡çÁÑ¡¨ÐäÁèÊÍ´¤Åéͧ¡Ñº·ÒÃìà¡çµ·Õ ¡Ó˹´ÁÒãËé
ËÁÒÂà˵Ø
㹡ÒÃËÒ·ÒÃìà¡çµ H ¨Ò¡ÊÁ¡Òà (3.20) áÅÐÍÕ¤ÇÍäÅà«ÍÃì
F
¨Ò¡ÊÁ¡Òà (3.21) ¢Ñ ¹µÍ¹
áá·Õ ¤ÇÃ·Ó ¤×Í ¡ÒÃËÒ¤èÒ àÁ·ÃÔ¡«ì R, A, áÅÐ P «Ö § àÁ·ÃÔ¡«ì ·Ñ § ÊÒÁ¹Õ ÊÒÁÒöËÒä´é §èÒÂã¹·Ò§ »¯ÔºÑµÔ ¡ÅèÒǤ×Í ¨Ò¡áºº¨ÓÅͧ·Õ ãªé 㹡ÒÃÍ͡Ẻ·ÒÃìà¡çµ µÒÁÃÙ» ·Õ 3.2 ãËé ·Ó¡ÒÃÊè§ ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ
{ak }
¨Ó¹Ç¹Ë¹Ö §à«¡àµÍÃì (sector) ËÃ×Í 4096 ºÔµ à¢éÒä»ã¹Ãкº à¾× Í·Ó¡ÒÃà¢Õ¹¢éÍÁÙÅŧ
ä»ã¹Ê× ÍºÑ¹·Ö¡ (¹Ñ ¹¤×Í ¼Ùéãªé·ÃÒºá¹è¹Í¹ÇèÒ
{ak }
¤×ÍÍÐäÃ) ¨Ò¡¹Ñ ¹ ¡çãËéËÑÇÍèÒ¹·Ó¡ÒÃÍèÒ¹¢éÍÁÙÅ
¨Ò¡Ê× Í ºÑ¹·Ö¡ áÅéÇ Êè§ ÊÑÒ³ read back ·Õ ä´é ¼èÒ¹ä»Âѧ ǧ¨Ã¡Ãͧ¼èÒ¹µ Ó áÅÐǧ¨ÃªÑ¡ µÑÇÍÂèÒ§ ·Ó ãËéä´é¼ÅÅѾ¸ìà» ¹ÅӴѺ¢éÍÁÙÅ à¾ÃÒÐ©Ð¹Ñ ¹ àÁ× Íä´é¢éÍÁÙÅ
{sk }
{ak }
«Ö §ã¹·Ò§»¯ÔºÑµÔ ¼ÙéãªéÊÒÁÒö·Õ ¨Ð·ÃÒºä´éÇèÒ¢éÍÁÙÅ
áÅÐ
{sk }
{sk }
¤×ÍÍÐäÃ
áÅéÇ ¡ç·Ó¡ÒäӹdzËÒ¤èÒàÁ·ÃÔ¡«ì R, A, áÅÐ P â´Â
ãªéâ»Ãá¡ÃÁ MATLAB [13] ËÃ×Í SCILAB [14] ¨ÐàËç¹ä´éÇèÒ ÇÔ¸Õ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµáºº MMSE ÊÒÁÒö·Óä´é§èÒÂã¹·Ò§»¯ÔºÑµÔ àÁ× Íà·Õº¡ÑºÇÔ¸Õ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµáººÍ× ¹æ
3.3
¼Å¡Ò÷´Åͧ
ã¹Êèǹ¹Õ ¨Ð·Ó¡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº·Õ ãªé ·ÒÃìà¡çµ ẺµèÒ§æ â´Âãªé Ẻ¨ÓÅͧã¹ÃÙ» 3
·Õ 3.2 ÊÓËÃѺ Ãкº¡Òúѹ·Ö¡ Ẻá¹ÇµÑ §
¨Ò¡ÃÙ» ·Õ 3.2 ÊÑÒ³ read back ·Õ ä´é ¨Ò¡ËÑÇÍèÒ¹
ÊÒÁÒöà¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§ÊÁ¡Òä³ÔµÈÒʵÃìä´é ¤×Í [40]
p(t) =
S−1 X
bk g(t − kT + ∆tk ) + n(t)
(3.26)
k=0
3
¡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº·Õ ãªé ·ÒÃìà¡çµ ẺµèÒ§æ ÊÓËÃѺ Ãкº¡Òúѹ·Ö¡ Ẻá¹Ç¹Í¹ ÊÒÁÒöÈÖ¡ÉÒ
ÃÒÂÅÐàÍÕ´ä´éã¹ [19]
3.3.
¼Å¡Ò÷´Åͧ
55
â´Â·Õ
bk = (ak − ak−1 )/2
ʶҹкǡËÃ×Íź áÅÐ ºÔµµÑÇ·Õ
k
¤×Í ºÔµà»ÅÕ Â¹Ê¶Ò¹Ð (àÁ× Í
bk = 0
∆t
¡Òúѹ·Ö¡ Ẻá¹ÇµÑ § µÒÁÊÁ¡ÒÃ·Õ (1.2),
n(t)
ÊÍ´¤Åéͧ¡Ñº¡ÒÃà»ÅÕ Â¹á»Å§
ËÁÒ¶֧ äÁèÁÕ¡ÒÃà»ÅÕ Â¹á»Å§Ê¶Ò¹Ð),
·Õ Áըӹǹ·Ñ §ËÁ´ 4096 ºÔµ (1 à«¡àµÍÃì),
a jitter noise), áÅÐ
bk = ±1
ak ∈ ±1
¤×Í ¢éÍÁÙÅÍÔ¹¾Øµ
g(t) ¤×Í ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹Ð¢Í§Ãкº
¤×Í ÊÑҳú¡Ç¹¨Ôµ àµÍÃì ¢Í§Ê× Í ºÑ¹·Ö¡ (medi
¤×Í ÊÑҳú¡Ç¹à¡ÒÊì ÊÕ ¢ÒÇẺºÇ¡ (AWGN) ·Õ ÁÕ ¤ÇÒÁ˹Òá¹è¹
Ê໡µÃÑÁ ¡ÓÅѧẺÊͧ´éÒ¹à·èÒ ¡Ñº
N0 /2
ã¹Ë¹Ñ§Ê×Í àÅèÁ ¹Õ ÊÑҳú¡Ç¹¨Ôµ àµÍÃì ¢Í§Ê× Í ºÑ¹·Ö¡
¨Ð¶Ù¡ ¨ÓÅͧãËé ÁÕ ÅѡɳÐà» ¹ ¡ÒÃàÅ× Í¹µÓáË¹è§ ¢Í§¡ÒÃà»ÅÕ Â¹Ê¶Ò¹ÐẺÊØèÁ (random transi tion shift) «Ö §ÁÕ¿ §¡ìªÑ¹¤ÇÒÁ˹Òá¹è¹¤ÇÒÁ¹èÒ¨Ðà» ¹áººà¡ÒÊìà«Õ¹ (Gaussian probability den sity function) ·Õ ÁÕ ¤èÒà©ÅÕ Â à·èÒ ¡Ñº ¤èÒ ÈÙ¹Âì áÅФèÒ ¤ÇÒÁá»Ã»Ãǹà·èÒ ¡Ñº
N (0, |bk |σj2 ))
áÅж١¨Ó¡Ñ´ãËéÁÕ¤èÒäÁèà¡Ô¹
T
¤×Í ¤èÒÊÑÁºÙóì¢Í§
à«ÅÅì
áÅÐ
|bk |
ÊÑÒ³ read back,
p(t),
T /2)
àÁ× Í
σj
|bk |σj2
(¹Ñ ¹¤×Í
∆tk ∼
¨Ð¶Ù¡¡Ó˹´à» ¹¨Ó¹Ç¹à»ÍÃìà«ç¹µì¢Í§ºÔµ
bk
¨Ð¶Ù¡ Êè§ ¼èÒ¹ä»Âѧ ǧ¨Ã¡Ãͧ¼èÒ¹µ Ó ºÑµà·ÍÃìàÇÔÃìµ (Butterworth)
Íѹ´Ñº·Õ 7 áÅШзӡÒêѡµÑÇÍÂèÒ§ÊÑÒ³
p(t)
´éÇÂÍѵÃÒ¤ÇÒÁ¶Õ à·èҡѺ
1/T
(¹Ñ ¹¤×Í ¢éÍÁÙÅá«Á
à» ÅáµèÅÐá«Áà» Å·Õ ä´é¨Ò¡¡ÒêѡµÑÇÍÂèÒ§¨ÐÍÂÙèËèÒ§¡Ñ¹ 1 ºÔµà«ÅÅì) â´Âã¹·Õ ¹Õ ¨ÐÊÁÁصÔÇèÒ ¡Ãкǹ¡Òà 4
à¢éÒ ¨Ñ§ËÇÐÃÐËÇèÒ§ÊÑÒ³ read back áÅÐǧ¨ÃªÑ¡ µÑÇÍÂèÒ§ à» ¹ ẺÊÁºÙóì nization) ¨Ò¡¹Ñ ¹ ÅӴѺ¢éÍÁÙÅ
{sk }
(perfect synchro
¡ç¨Ð¶Ù¡Êè§ä»ÂѧÍÕ¤ÇÍäÅà«ÍÃì à¾× Í»ÃѺÃÙ»ÃèÒ§ÊÑÒ³ãËéà» ¹ä»
µÒÁ·ÒÃìà¡çµ ·Õ Ãкºµéͧ¡Òà áÅÐǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¡ç ¨Ð·Ó¡ÒöʹÃËÑÊ ÅӴѺ ¢éÍÁÙÅ ÅӴѺ¢éÍÁÙÅ
{ak }
{yk }
à¾× Í ËÒ
·Õ à» ¹ä»ä´éÁÒ¡·Õ ÊØ´
㹡Ò÷´Åͧ ¨Ð¹ÔÂÒÁ Electronics SNR (ËÃ×ÍàÃÕ¡ÊÑ ¹æ ÇèÒ SNR) ãËéÁÕ¤èÒà·èҡѺ
µ SNR = 10 log10 â´Â·Õ
Vp = g(∞) = 1
pulse) ³ àÇÅÒ
t
=
∞
Vp 2 σ2
¶ (dB)
¤×Í ¢¹Ò´¢Í§ÊÑÒ³¾ÑÅÊì à»ÅÕ Â¹Ê¶Ò¹ÐàÍ¡à·È áÅÐ
σn2
=
N0 /(2T )
(3.27)
(isolated transition
¤×Í ¡ÓÅѧ ¢Í§ÊÑҳú¡Ç¹
n(t)
¹Í¡¨Ò¡¹Õ
áµèÅШش ¢Í§ BER ¶Ù¡ ¤Ó¹Ç³â´Âãªé ¢éÍÁÙÅ ËÅÒÂæ à«¡àµÍÃì ¨¹¡ÇèÒ ¨Ðä´é ¢éͼԴ¾ÅÒ´ÃÇÁäÁè ¹éÍ¡ÇèÒ 1000 ºÔµ áÅзÒÃìà¡çµ áÅÐÍÕ¤ÇÍäÅà«ÍÃì ·Õ ãªé ¨Ð¶Ù¡ Í͡ẺãËé àËÁÒÐÊÁ¡Ñº ¡Ò÷ӧҹã¹áµèÅÐ ND 4
¹Ñ ¹¤×Í Ç§¨ÃªÑ¡µÑÇÍÂèÒ§·ÃÒºÇèÒ ¨Ð·Ó¡ÒêѡµÑÇÍÂèÒ§ÊÑÒ³ read back ·Õ µÓá˹è§ã´ à¾× ÍãËé¢éÍÁÙÅá«Áà» Å·Õ ä´éÍÍ¡
ÁÒÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡·Õ ÊØ´
56
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
³ ÃдѺ BER =
µÑÇÍÂèÒ§·Õ 3.1
10−5
µÑÇÍÂèÒ§¹Õ ¨ÐáÊ´§¢Ñ ¹µÍ¹¡ÒäӹdzËÒ·ÒÃìà¡çµáÅÐÍÕ¤ÇÍäÅà«ÍÃì µÒÁÊÁ¡Òà (3.19)
(3.21) ãËé¾Ô¨ÒóÒẺ¨ÓÅͧÃкº¡Òúѹ·Ö¡áººá¹ÇµÑ §ã¹ÃÙ»·Õ 3.2 ÊÓËÃѺ ND = 1.5 áÅÐ S NR = 20 dB ¶éÒ¡Ó˹´ãËéÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ ¢Òà¢éҢͧÍÕ¤ÇÍäÅà«ÍÃì ·Õ ÊÍ´¤Åéͧ¡Ñº
{ak }
{ak }
¤×Í
{1,
=
{sk }
=
1, 1,
−1,
1,
−1}
áÅÐÅӴѺ¢éÍÁÙÅ´éÒ¹
{−0.5337, −0.0662,
0.8821, 0.8122,
0.3219, 0.0260} ¨§¤Ó¹Ç³ËÒ·ÒÃìà¡çµáºº GPR ¢¹Ò´ 3 á·ç» áÅÐÍÕ¤ÇÍäÅà«ÍÃ좹Ҵ 5 á·ç» ·Õ ÊÍ´¤Åéͧ¡Ñº·ÒÃìà¡çµ µÒÁà§× ͹䢺ѧ¤Ñº ´Ñ§µèÍä»¹Õ ¡) à§× ͹䢺ѧ¤ÑºáººâÁ¹Ô¡
¢) à§× ͹䢺ѧ¤Ñºáºº
h0 = 1
h1 = 1
¤) à§× ͹䢺ѧ¤Ñºáºº¾Åѧ§Ò¹Ë¹Ö §Ë¹èÇÂ
§) à§× ͹䢺ѧ¤Ñºáºº·ÒÃìà¡çµà©¾ÒÐ àÁ× Í¡Ó˹´ãËé ÇÔ¸Õ ·Ó
H(D) = 1 + 2D + D2
¢Ñ ¹µÍ¹áá㹡ÒÃËÒ·ÒÃìà¡çµ áÅÐÍÕ¤ÇÍäÅà«ÍÃì ¤×Í ¡ÒÃËÒ¤èÒ àÁ·ÃÔ¡«ì R, A, áÅÐ P µÒÁ
ÊÁ¡Òà (3.6) (3.8) àÁ× Í
S = 6, K = 2,
áÅÐ
L=3
«Ö § ¨Ò¡¡Òäӹdz¨Ðä´éÇèÒ àÁ·ÃÔ¡«ì
ÁÕ¤èÒà·èҡѺ
0.3052
0.1926
−0.0549 −0.1440 −0.0868
1.0000
−0.2000
0.1926 0.3052 0.1926 −0.0549 −0.1440 R = −0.0549 0.1926 0.3052 0.1926 −0.0549 −0.1440 −0.0549 0.1926 0.3052 0.1926 −0.0868 −0.1440 −0.0549 0.1926 0.3052 àÁ·ÃÔ¡«ì
A
ÁÕ¤èÒà·èҡѺ
0.5000
A = −0.2000 1.0000 −0.2000 0.5000 −0.2000 1.0000
R
3.3.
¼Å¡Ò÷´Åͧ
áÅÐàÁ·ÃÔ¡«ì
P
57
ÁÕ¤èÒà·èҡѺ
0.4976
0.3867
0.1740
0.2664 0.4976 0.3867 P = −0.0390 0.2664 0.4976 −0.1983 −0.0390 0.2664 −0.0994 −0.1983 −0.0390 â´Âã¹·Õ ¹Õ ¨Ð¢ÍáÊ´§µÑÇÍÂèÒ§¡ÒäӹdzËÒÊÁÒªÔ¡ ËÒä´é¨Ò¡
A(i, j) = E
" 5 X
(i, j)
ºÒ§µÑÇ ¢Í§àÁ·ÃÔ¡«ì A ´Ñ§¹Õ àÁ·ÃÔ¡«ì A
# ak−i ak−j , 0 ≤ i, j ≤ 2
k=0
{ak } = {a0 , a1 , a2 , a3 , a4 , a5 } # " 5 X ak ak A(0, 0) = E
à¹× ͧ¨Ò¡ ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
=
{1,
1, 1,
−1,
1,
−1}
´Ñ§¹Ñ ¹
k=0
= E [a0 a0 + a1 a1 + a2 a2 + a3 a3 + a4 a4 + a5 a5 ] 1 = [(1)(1) + (1)(1) + (1)(1) + (−1)(−1) + (1)(1) + (−1)(−1)] 6 = 1 A(0, 1) = E
" 5 X
# ak ak−1
k=0
= E [a1 a0 + a2 a1 + a3 a2 + a4 a3 + a5 a4 ] 1 = [(1)(1) + (1)(1) + (−1)(1) + (1)(−1) + (−1)(1)] 5 = −0.2 A(0, 2) = E
" 5 X
# ak ak−2
k=0
= E [a2 a0 + a3 a1 + a4 a2 + a5 a3 ] 1 = [(1)(1) + (−1)(1) + (1)(1) + (−1)(−1)] 4 = 0.5
58
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ÊÓËÃѺ ÊÁÒªÔ¡ ¢Í§àÁ·ÃÔ¡«ìÍ× ¹æ ¡ç ÊÒÁÒö¤Ó¹Ç³ËÒä´é ã¹ÅѡɳÐà´ÕÂǡѹ àÁ× Í ä´é àÁ·ÃÔ¡«ì R, A, áÅÐ P áÅéÇ áÅШҡ⨷Âì¨Ðä´éÇèÒ
H
=
[h0 , h1 , h2 ]T
áÅÐ
F
=
[f−2 , f−1 , f0 , f1 , f2 ]T
à¾ÃÒÐ©Ð¹Ñ ¹
·ÒÃìà¡çµáÅÐÍÕ¤ÇÍäÅà«ÍÃìÊÒÁÒöËÒ¤èÒä´éµÒÁÊÁ¡Òà (3.19) (3.21) µÒÁà§× ͹䢺ѧ¤ÑºµèÒ§æ ´Ñ§¹Õ
¡) à§× ͹䢺ѧ¤ÑºáººâÁ¹Ô¡
−0.8587]T ,
áÅÐ
¢) à§× ͹䢺ѧ¤Ñº Ẻ
−1.6613]T ,
áÅÐ
F
=
h0 = 1
=
=
[1, 0, 0]T , λ = −0.0837, H
¨Ðä´é ÇèÒ
J
[0, 1, 0]T , λ = −2.2814, H
=
=
[4.7796, 1,
[7.0507, 1.8111, −4.6115, 3.0835, −1.9107]T
¤) à§× ͹䢺ѧ¤Ñºáºº¾Åѧ§Ò¹Ë¹Ö §Ë¹èÇ ¨Ðä´éÇèÒ¤èÒÅѡɳÐ੾ÒТͧàÁ·ÃÔ¡«ì ¤×Í
[1, 0.1754,
=
[1.0936, −0.0722, −0.9953, −0.2400, −0.0805]T
h1 = 1 F
I
¨Ðä´éÇèÒ
{1.5182, −0.0476, −1.0211}, H
=
[0.3843, 0.7335, 0.5607]T ,
(A − PT R−1 P)
áÅÐ
F
=
[1.7138,
1.1288, 0.0256, 1.5285, −0.6140]T §) à§× ͹䢺ѧ¤Ñºáºº·ÒÃìà¡çµà©¾ÒÐ àÁ× Í
H
=
[1, 2, 1]T
¨Ðä´éÇèÒ
F
=
[4.2431, 2.6064, 0.1106,
3.1983, −1.3164]T
㹡Ò÷´Åͧ¹Õ ¨Ð·Ó¡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº·Õ ãªé ·ÒÃìà¡çµ ẺµèÒ§æ ·Ñ § 5 Ẻ ¤×Í
1) ·ÒÃìà¡çµáºº PR1 ¹Ñ ¹¤×Í
2) ·ÒÃìà¡çµ PR2 ¹Ñ ¹¤×Í
H(D) = 1 + D
ËÃ×Íà¢Õ¹᷹´éÇÂ
H(D) = 1 + 2D + D2
3) ·ÒÃìà¡çµ·Õ Í͡ẺµÒÁà§× ͹䢺ѧ¤ÑºáººâÁ¹Ô¡
4) ·ÒÃìà¡çµ·Õ Í͡ẺµÒÁà§× ͹䢺ѧ¤Ñºáºº
[1, 1]
ËÃ×Íà¢Õ¹᷹´éÇÂ
[1, 2, 1]
h0 = 1
h1 = 1
5) ·ÒÃìà¡çµ·Õ Í͡ẺµÒÁà§× ͹䢺ѧ¤Ñºáºº¾Åѧ§Ò¹Ë¹Ö §Ë¹èÇÂ
ÃÙ» ·Õ 3.3 à»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº·Õ ãªé ·ÒÃìà¡çµµèÒ§æ ÊÓËÃѺ ND = 2 â´Â·Õ Ãкº¨Ðãªé ·ÒÃìà¡çµáºº GPR ·Õ Áըӹǹ 3 á·ç» «Ö Ö§ã¹·Õ ¹Õ ¨Ðä´éÇèÒ ·ÒÃìà¡çµáººâÁ¹Ô¡
h0 = 1
¤×Í
1 + 1.15D
3.3.
¼Å¡Ò÷´Åͧ
59
−1
10
PR1 [1 1] PR2 [1 2 1] h0 = 1 h1 = 1 Unit energy
−2
BER
10
−3
10
−4
10
−5
10 14
15
16
17
18
19
20
SNR (dB)
ÃÙ»·Õ 3.3: »ÃÐÊÔ·¸ÔÀÒ¾ã¹ÃÙ» BER ¢Í§Ãкº·Õ ãªé·ÒÃìà¡çµµèÒ§æ ÊÓËÃѺ ND = 2
+
0.48D2 ,
·ÒÃìà¡çµáºº
0.45 + 0.77D
+
h1 = 1
0.45D2
¤×Í
0.55 + D
+
0.55D2 ,
áÅзÒÃìà¡çµáºº¾Åѧ§Ò¹Ë¹Ö §Ë¹èÇ ¤×Í
¨Ò¡ÃÙ»·Õ 3.3 ¨ÐàËç¹ä´éÇèÒ â´ÂÀÒ¾ÃÇÁáÅéÇÃкº·Õ ãªé·ÒÃìà¡çµáºº GPR
¨ÐÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒÃкº·Õ ãªé·ÒÃìà¡çµáºº PR áÅÐÃкº·Õ ãªé·ÒÃìà¡çµáººâÁ¹Ô¡¨ÐãËé»ÃÐÊÔ·¸ÔÀÒ¾ ´Õ·Õ ÊØ´ â´Â੾ÒÐÍÂèÒ§ÂÔ §·Õ ND ÊÙ§æ à¾× Íà» ¹¡ÒÃÂ×¹ÂѹÇèÒ ·ÒÃìà¡çµáºº GPR ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒ·ÒÃìà¡çµáºº PR ¨Ð·Ó¡ÒÃà»ÃÕº à·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº·Õ ND = 2.5 áµè ¤ÃÒÇ¹Õ ¨Ðãªé ·ÒÃìà¡çµ Ẻ GPR ·Õ ÁÕ ¨Ó¹Ç¹ 5 á·ç» (¨Ó¹Ç¹á·ç»ÁÒ¡¢Ö ¹ à¾× ÍÃͧÃѺ ND ·Õ à¾Ô Á¢Ö ¹) «Ö §ã¹¡Ã³Õ¹Õ ¨Ðä´éÇèÒ ·ÒÃìà¡çµáººâÁ¹Ô¡
1 + 1.42D + 1.06D2 + 0.43D3 + 0.08D4 , 0.53D3 + 0.14D4 , 0.2D4
·ÒÃìà¡çµáºº
áÅзÒÃìà¡çµáºº¾Åѧ§Ò¹Ë¹Ö §Ë¹èÇ ¤×Í
h1 = 1
¤×Í
h0 = 1
¤×Í
0.47 + D + 0.96D2 +
0.20 + 0.5D + 0.66D2 + 0.5D3 +
ÃÙ» ·Õ 3.4 à»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº·Õ ãªé ·ÒÃìà¡çµµèÒ§æ ·Õ ND = 2.5 ¨Ò¡¼ÅÅѾ¸ì ·Õ
ä´é¨Ð¾ºÇèÒ Ãкº·Õ ãªé·ÒÃìà¡çµáºº
h0 = 1
¨ÐãËé»ÃÐÊÔ·¸ÔÀÒ¾´Õ·Õ ÊØ´ ·Ñ §¹Õ à» ¹à¾ÃÒÐÇèÒ ·ÒÃìà¡çµáºº
GPR (â´Â੾ÒÐÍÂèÒ§ÂÔ § ·ÒÃìà¡çµáººâÁ¹Ô¡
h0 = 1)
¨ÐÁÕ ¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ã¡Åéà¤Õ§¡Ñº ¼Å
60
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
−1
10
PR1 [1 1] PR2 [1 2 1] h0 = 1 h1 = 1 Unit energy
−2
BER
10
−3
10
−4
10
−5
10 19.0
19.5
20.0
20.5
21.0
21.5
22.0
SNR (dB)
ÃÙ»·Õ 3.4: »ÃÐÊÔ·¸ÔÀÒ¾ã¹ÃÙ» BER ¢Í§Ãкº·Õ ãªé·ÒÃìà¡çµµèÒ§æ ·Õ ND = 2.5
µÍºÊ¹Í§¢Í§ªèͧÊÑÒ³ÁÒ¡¡ÇèÒ·ÒÃìà¡çµáºº PR ´Ñ§áÊ´§ã¹ÃÙ»·Õ 3.5 ¹Í¡¨Ò¡¹Õ ·ÒÃìà¡çµ Ẻ GPR ÁÕ á¹Çâ¹éÁ ·Õ ¨ÐªèÇ·ÓãËé ͧ¤ì»ÃСͺ¢Í§ÊÑҳú¡Ç¹·Õ ´éÒ¹ ¢Òà¢éҢͧǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔÁÕÅѡɳÐà» ¹ÊÑҳú¡Ç¹ÊÕ¢ÒÇ (white noise) «Ö §à» ¹¢éÍ¡Ó˹´ ËÅÑ¡·Õ ¨ÐªèÇ·ÓãËéǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔÊÒÁÒö·Ó§Ò¹ä´éÍÂèÒ§ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡·Õ ÊØ´ [15] ¡ÒõÃǨ ÊͺÊÑҳú¡Ç¹
wk
ÇèÒÁդسÊÁºÑµÔà» ¹ÊÑҳú¡Ç¹ÊÕ¢ÒÇËÃ×ÍäÁè ·Óä´éâ´Â¡ÒÃËÒ¤èÒÍѵÊË
ÊÑÁ¾Ñ¹¸ì (auto correlation) ¢Í§ÅӴѺ¢éÍÁÙÅ
{wk }
(´ÙÃÙ»·Õ 3.2) ¶éÒ¼ÅÅѾ¸ì·Õ ä´éÁÕ¤èÒ੾ÒеÓá˹è§
·Õ ¼ÅµèÒ§¢Í§àÇÅÒÁÕ¤èÒà·èҡѺ¤èÒÈÙ¹Âì áÅÐÁÕ¤èÒÈÙ¹Âì (ËÃ×ͤèÒã¡ÅéÈÙ¹Âì) àÁ× Í¼ÅµèÒ§¢Í§àÇÅÒÁÕ¤èÒà» ¹¤èÒ Í× ¹æ ¨ÐÊÃØ» ä´é ÇèÒ ÅӴѺ ¢éÍÁÙÅ ÊÑÁ¾Ñ¹¸ì ¢Í§ÅӴѺ ¢éÍÁÙÅ
{wk }
Ãкº·Õ ãªé·ÒÃìà¡çµáººâÁ¹Ô¡
wk
ÁÕ ÅѡɳÐà» ¹ ÊÑҳú¡Ç¹ÊÕ ¢ÒÇ ÃÙ» ·Õ 3.6 áÊ´§¤èÒ ÍѵÊË
ÊÓËÃѺ Ãкº·Õ ãªé ·ÒÃìà¡çµ ẺµèÒ§æ ·Õ ND = 2.5 ¨Ò¡ÃÙ» ¨Ð¾ºÇèÒ
h0 = 1
¨ÐÁռŷÓãËéͧ¤ì»ÃСͺ¢Í§ÊÑҳú¡Ç¹·Õ ´éÒ¹¢Òà¢éҢͧ
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÁÕÅѡɳÐà» ¹ÊÑҳú¡Ç¹ÊÕ¢ÒÇÁÒ¡·Õ ÊØ´ àÁ× Íà·Õº¡Ñº·ÒÃìà¡çµáººÍ× ¹æ à¹× ͧ¨Ò¡ ·ÒÃìà¡çµáººâÁ¹Ô¡
h0 = 1
ãËé »ÃÐÊÔ·¸ÔÀÒ¾´Õ ·Õ ÊØ´ àÁ× Í à»ÃÕºà·Õº¡Ñº ·ÒÃìà¡çµ Ẻ
3.3.
¼Å¡Ò÷´Åͧ
61
1.0
Channel ND=2.5 PR1 [1 1] PR2 [1 2 1] h0 = 1 h1 = 1
0.9
Normalized magnitude
0.8 0.7
Unit energy 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
Normalized frequency (fT)
ÃÙ»·Õ 3.5: ¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ ¢Í§·ÒÃìà¡çµáººµèÒ§æ à·Õº¡ÑºªèͧÊÑÒ³·Õ ND = 2.5
GPR Í× ¹æ ã¹ÊèǹµèÍä»¹Õ ÊÑÅѡɳì GPRn ¨Ð¶Ù¡ãªéá·¹·ÒÃìà¡çµáººâÁ¹Ô¡
n
h0 = 1
·Õ Áըӹǹ
á·ç» ÃÙ» ·Õ 3.7(a) à»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§·ÒÃìà¡çµ ẺµèÒ§æ ·Õ ND µèÒ§æ ÊÓËÃѺ ¡Ã³Õ ·Õ
äÁèÁÕÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡ (σj /T µéͧ¡ÒÃà¾× Í·Õ ¨Ð·ÓãËéä´é BER =
10−4
= 0%)
â´Â·Õ àÊé¹á¡¹
y
áÊ´§ SNR ·Õ Ãкº
´Ñ§¹Ñ ¹ Ãкº·Õ ãªé SNR ¹éÍ¡ÇèÒ¡çáÊ´§ÇèÒÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ
¡ÇèÒ ¨Ò¡ÃÙ»¨ÐàËç¹ä´éÇèÒ ·ÒÃìà¡çµáºº GPR ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒ·ÒÃìà¡çµáºº PR â´Â੾ÒÐÍÂèÒ§ÂÔ § ·Õ ND ÊÙ§æ ·Ñ §¹Õ à» ¹à¾ÃÒÐÇèÒ ·ÒÃìà¡çµáºº GPR ÁռŵͺʹͧàªÔ§¤ÇÒÁ¶Õ ã¡Åéà¤Õ§¡ÑºªèͧÊÑÒ³ ÁÒ¡¡ÇèÒ·ÒÃìà¡çµáºº PR (´ÙÃÙ»·Õ 3.5) ÃÙ»·Õ 3.7(b) à»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§·ÒÃìà¡çµáººµèÒ§æ ·Õ ÃдѺ¤ÇÒÁÃعáç¢Í§
σj /T
µèÒ§æ
ÊÓËÃѺ ND = 2.5 㹷ӹͧà´ÕÂǡѹ¨Ðä´éÇèÒ ·ÒÃìà¡çµáºº GPR ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒ (µéͧ¡Òà S NR ¹éÍ¡ÇèÒ à¾× ÍãËéä´é BER =
10−4
à·èҡѹ) ·ÒÃìà¡çµáºº PR ³ ·Ø¡ÃдѺ¤ÇÒÁÃعáç¢Í§
σj /T
Êѧࡵ¨Ð¾ºÇèÒ ·ÒÃìà¡çµáºº PR2 µéͧ¡Òà SNR ¹éÍ¡ÇèÒ (ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒ) ·ÒÃìà¡çµáºº EPR2 áÅÐ EEPR2 àÁ× Í ¤ÇÒÁÃعáç¢Í§
σj /T
ÁÕ ¤èÒ ÁÒ¡ ·Ñ §¹Õ ÍÒ¨¨Ðà» ¹ à¾ÃÒÐÇèÒ ·ÒÃìà¡çµ Ẻ PR ·Õ ÁÕ
62
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
1.2
PR1 [1 1] PR2 [1 2 1] Monic (h0 = 1) h1 = 1 Unit energy
1.0
Normalized amplitude
0.8
0.6
0.4
0.2
0.0
−0.2
−0.4
−0.6 0
1
2
3
4
5
6
7
8
9
Time difference (in bit period)
ÃÙ»·Õ 3.6: ÍѵÊËÊÑÁ¾Ñ¹¸ì¢Í§ÅӴѺ¢éÍÁÙÅ
{wk }
ÊÓËÃѺÃкº·Õ ·ÒÃìà¡çµáººµèÒ§æ ·Õ ND = 2.5
¨Ó¹Ç¹á·ç» ¹éͨÐÁÕ ¤ÇÒÁÍè͹äËÇ (sensitive) ¡Ñº ÊÑҳú¡Ç¹¨Ôµ àµÍÃì ¢Í§Ê× Í ºÑ¹·Ö¡ ¹éÍ¡ÇèÒ ·ÒÃìà¡çµ Ẻ PR ·Õ ÁÕ ¨Ó¹Ç¹á·ç» ÁÒ¡ ã¹·Ò§µÃ§¡Ñ¹¢éÒÁ ·ÒÃìà¡çµáºº GPR ¨ÐãËé »ÃÐÊÔ·¸ÔÀÒ¾·Õ ´Õ ¢Ö ¹àÊÁÍ àÁ× Í ·ÒÃìà¡çµ·Õ ãªéÁըӹǹá·ç»ÁÒ¡¢Ö ¹ â´ÂäÁè¤Ó¹Ö§¶Ö§ÃдѺ¤ÇÒÁÃعáç¢Í§
3.4
σj /T
ÊÃØ»·éÒº·
·ÒÃìà¡çµáºº GPR ãËé »ÃÐÊÔ·¸ÔÀÒ¾´Õ ¡ÇèÒ ·ÒÃìà¡çµ Ẻ PR â´Â੾ÒÐÍÂèÒ§ÂÔ § ·Õ ND ÊÙ§æ ·Ñ §¹Õ à» ¹ à¾ÃÒÐÇèÒ ·ÒÃìà¡çµáºº GPR ÁÕ ¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ã¡Åéà¤Õ§¡Ñº ¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§ªèͧ ÊÑÒ³ÁÒ¡¡ÇèÒ ·ÒÃìà¡çµ Ẻ PR ¹Í¡¨Ò¡¹Õ ·ÒÃìà¡çµáºº GPR Âѧ ÁÕ á¹Çâ¹éÁ ·Õ ¨ÐªèÇ·ÓãËé ͧ¤ì »ÃСͺ¢Í§ÊÑҳú¡Ç¹·Õ ´éÒ¹¢Òà¢éÒ ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÁÕ ÅѡɳÐà» ¹ ÊÑҳú¡Ç¹ÊÕ ¢ÒÇ «Ö §Ê觼ŷÓãËéǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔÊÒÁÒö·Ó§Ò¹ä´éÍÂèÒ§ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡ÂÔ §¢Ö ¹ ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ áÅÐÍÕ¤ÇÍäÅà«ÍÃì ÊÒÁÒö·Óä´é ËÅÒÂÇÔ¸Õ¡Òà áµè ã¹·Ò§»¯ÔºÑµÔ ¨Ð¾ºÇèÒ¡ÒÃ
3.4.
ÊÃØ»·éÒº·
63
ND 2 2.5 3
25
SNR required to achieve BER = 10
−4
in dB
26
5−tap GPR targets [1 1.14 0.58 0.16 0.03] [1 1.34 0.99 0.43 0.09] [1 1.44 1.31 0.74 0.22]
24
23
22
PR2 [1 2 1] EPR2 [1 3 3 1] EEPR2 [1 4 6 4 1] GPR5
21
20
2
2.1
2.2
2.3
(a)
2.4
2.5
2.6
2.7
2.8
2.9
3
Normalized density (ND)
SNR required to achieve BER = 10−4 in dB
30
Jitter (%) 0 3 6 9
29
28
[1 [1 [1 [1
5−tap GPR targets 1.34 0.99 0.43 0.09] 1.33 0.94 0.36 0.06] 1.27 0.72 0.13 −0.03] 1.02 0.15 −0.16 −0.01]
27
26
25
24
23
PR2 [1 2 1] EPR2 [1 3 3 1] EEPR2 [1 4 6 4 1] GPR5
22
21
20
0
1
2
3
4
(b)
ÃÙ»·Õ 3.7: (a) ¡ÃÒ¿ÃÐËÇèÒ§ SNR ·Õ µéͧ¡Òà áÅÐ ND àÁ× Í ·Õ µéͧ¡Òà áÅÐ
σj
·Õ ND = 2.5
5
6
7
8
Jitter (%)
σj = 0%
áÅÐ (b) ¡ÃÒ¿ÃÐËÇèÒ§ SNR
64
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
Í͡Ẻ·ÒÃìà¡çµ ´éÇÂÇÔ¸Õ¡Òà MMSE ÊÒÁÒö·Óä´é §èÒ¡ÇèÒ ÇÔ¸Õ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ ẺÍ× ¹æ áÅÐ à¹× ͧ¨Ò¡¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ ´éÇÂÇÔ¸Õ¡Òà MMSE ÁÕ à§× ͹䢺ѧ¤Ñº ËÅÒÂẺ áµè ¨Ò¡¡Ò÷´Åͧ¾º ÇèÒ ·ÒÃìà¡çµ·Õ Í͡Ẻâ´Âà§× ͹䢺ѧ¤Ñº ẺâÁ¹Ô¡ ¨ÐãËé »ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡¡ÇèÒ ·ÒÃìà¡çµ ·Õ Í͡Ẻâ´Â à§× ͹䢺ѧ¤ÑºÍ× ¹ ¹Í¡¨Ò¡¹Õ ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ·Õ ´Õ¡ç¤ÇÃ·Õ ¨ÐÍ͡Ẻ·ÒÃìà¡çµãËéàËÁÒÐÊÁ¡ÑºÊÀÒ¾ áÇ´ÅéÍÁ㹡Ò÷ӧҹ¢Í§µÑǪԻªèͧÊÑÒ³ÍèÒ¹ (read channel chip) àªè¹ Í͡Ẻ·ÒÃìà¡çµÊÓËÃѺ áµèÅÐ ND, SNR, áÅÐ
σj /T
à» ¹µé¹ à¾× Í·Õ ¨Ð·ÓãËé ä´é ·ÒÃìà¡çµ ·Õ ÁÕ »ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡·Õ ÊØ´ÀÒÂãµé
ÊÀÒ¾áÇ´ÅéÍÁ¡Ò÷ӧҹ·Õ ¡Ó˹´
3.5
à຺½ ¡ËÑ´·éÒº·
1. ¨§¾ÔÊÙ¨¹ìÊÁ¡Òà (3.9)
2. ¨§¾ÔÊÙ¨¹ìÊÁ¡Òà (3.24)
3. ¾Ô¨ÒóÒẺ¨ÓÅͧµÒÁÃÙ»·Õ 3.2 áµèà» ¹Ãкº¡Òúѹ·Ö¡áººá¹Ç¹Í¹ (longitudinal record ing) ·Õ ND = 2 áÅÐ SNR = 20 dB ¶éÒ¡Ó˹´ãËéÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ 1, 1} áÅÐÅӴѺ¢éÍÁÙÅ´éÒ¹¢Òà¢éҢͧÍÕ¤ÇÍäÅà«ÍÃì·Õ ÊÍ´¤Åéͧ¡Ñº 0.6184,
−0.1537,
0.0469,
−0.2469,
{ak }
{ak }
¤×Í
=
{1, −1,
{sk }
=
1,
−1,
{0.0767,
0.2096} ¨§¤Ó¹Ç³ËÒ·ÒÃìà¡çµ Ẻ GPR ¢¹Ò´ 3
á·ç» áÅÐÍÕ¤ÇÍäÅà«ÍÃ좹Ҵ 5 á·ç» ·Õ ÊÍ´¤Åéͧ¡Ñº·ÒÃìà¡çµ µÒÁà§× ͹䢺ѧ¤ÑºµèÍ仹Õ
¡) à§× ͹䢺ѧ¤ÑºáººâÁ¹Ô¡ ¢) à§× ͹䢺ѧ¤Ñºáºº
h0 = 1
h1 = 1
¤) à§× ͹䢺ѧ¤Ñºáºº¾Åѧ§Ò¹Ë¹Ö §Ë¹èÇ §) à§× ͹䢺ѧ¤Ñºáºº·ÒÃìà¡çµà©¾ÒÐ àÁ× Í¡Ó˹´ãËé
H(D) = 1 − D2
4. ¨§à¢Õ¹â»Ãá¡ÃÁ´éÇÂÀÒÉÒ SCILAB ËÃ×Í MATLAB à¾× ÍÇÒ´ÃÙ»ÊÑÒ³¢Í§·ÒÃìà¡çµáºº [1
−4
1], [1
−2 −2
1], áÅÐ [1 0
−8
0 1] à¾× ;ÔÊÙ¨¹ìÇèÒ ·ÒÃìà¡çµàËÅèÒ¹Õ ÊÍ´¤Åéͧ¡Ñº¼Å
µÍºÊ¹Í§ä´ºÔµã¹â´àÁ¹àÇÅҢͧÃкº¡Òúѹ·Ö¡áººá¹ÇµÑ §
º··Õ 4
ǧ¨ÃµÃǨËÒ PRML
㹺·¹Õ ¨Ð͸ԺÒÂ¾× ¹°Ò¹¡Ò÷ӧҹ¢Í§à·¤¹Ô¤ ¼ÅµÍºÊ¹Í§ºÒ§Êèǹ¤ÇèÐà» ¹ ÁÒ¡ÊØ´ (PRML: partial response maximum likelihood) [27] «Ö §à» ¹à·¤¹Ô¤ËÅÑ¡·Õ ãªé㹡ÒõÃǨËÒ¢éÍÁÙŢͧÃкº ¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì â´Â¨Ðà¹é¹ ä»·Õ ËÅÑ¡¡Ò÷ӧҹ¢Í§ÍÑÅ¡ÍÃÔ·ÖÁ ÇÕà·ÍÃìºÔ (Viterbi algorithm) «Ö § ¶×Í ÇèÒ à» ¹ ÍÑÅ ¡ÍÃÔÖ ·ÖÁ ·Õ ãªé 㹡ÒöʹÃËÑÊ ¢éÍÁÙÅ ·Õ ÁÕ »ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡·Õ ãªé ã¹ ÎÒÃì´´ÔÊ¡ìä´Ã¿ì» ¨¨ØºÑ¹
4.1
º·¹Ó
à» ¹ ·Õ ·ÃÒº¡Ñ¹ ÇèÒ àÁ× Í ¤ÇÒÁ¨Ø ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì à¾Ô Á ¢Ö ¹ ¼Å¡Ãзº·Õ à¡Ô´ ¨Ò¡¡ÒÃá·Ã¡ÊÍ´ÃÐËÇèÒ§ ÊÑÅѡɳì (ISI) ¡ç¨ÐÂÔ §ÁÒ¡¢Ö ¹ ·ÓãËéäÁèÊÒÁÒöãªé§Ò¹Ç§¨ÃµÃǨËҨشÊÙ§ÊØ´ (peak detector) ä´éÍÕ¡ µèÍä» ´Ñ§¹Ñ ¹ à¾× ͨѴ¡ÒáѺ ISI ¨Ó¹Ç¹ÁÒ¡àËÅèÒ¹Õ Ç§¨ÃµÃǨËÒ PRML [27] ¨Ö§ä´é¶Ù¡¾Ñ²¹ÒáÅÐ ¹ÓÁÒãªé§Ò¹ã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì¨¹¶Ö§·Ø¡Çѹ¹Õ â´Â·Ñ Çä» ¤ÓÇèÒ PRML ËÁÒ¶֧ à·¤¹Ô¤¡ÒÃãªé§Ò¹ÃèÇÁ ¡Ñ¹ÃÐËÇèÒ§ÍÕ¤ÇÍäÅà«ÍÃìẺ PR áÅÐǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ µÒÁÃÙ»·Õ 4.1 «Ö §ÊÒÁÒöáºè§¢Ñ ¹µÍ¹¡Òà ·Ó§Ò¹ÍÍ¡à» ¹ 2 ¢Ñ ¹µÍ¹ ¤×Í
1) »ÃѺÃÙ»ÃèÒ§¢Í§ÊÑÒ³ãËéà» ¹ä»µÒÁ·ÒÃìà¡çµ·Õ µéͧ¡ÒÃ
2) ¶Í´ÃËÑÊ¢éÍÁÙÅâ´Âãªéǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ·Õ ÊÃéÒ§¨Ò¡·ÒÃìà¡çµ·Õ ¡Ó˹´äÇé 65
66
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
noise
ak
{± 1}
receiving filter
channel
detector
equalizer
âk
target response
ÃÙ»·Õ 4.1: ËÅÑ¡¡ÒÃ¾× ¹°Ò¹¢Í§à·¤¹Ô¤ PRML
A( D)
N (D)
C (D)
F ( D)
detector
 ( D )
ÃÙ»·Õ 4.2: Ẻ¨ÓÅͧªèͧÊÑÒ³·Õ äÁèµèÍà¹× ͧ·Ò§àÇÅÒẺÊÁÁÙÅ
¢éʹբͧ¡ÒÃãªéà·¤¹Ô¤ PRML ¤×Í Ãкº¨Ð༪ԡѺ¡ÒâÂÒÂÊÑҳú¡Ç¹ (noise enhancement) ·Õ µ Ó áÅФÇÒÁ«Ñº«é͹¢Í§Ãкº¨Ð¹éÍÂŧ ã¹ÊèǹµèÍä»¹Õ ¨Ð͸ԺÒÂËÅÑ¡¡Ò÷ӧҹ¢Í§ÍÕ¤ÇÍäÅà«ÍÃì Ẻ PR áÅÐÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔÍÂèÒ§ÅÐàÍÕ´
4.2
ÍÕ¤ÇÍäÅà«ÍÃì
¾Ô¨ÒóÒẺ¨ÓÅͧªèͧÊÑÒ³·Õ äÁè µèÍà¹× ͧ·Ò§àÇÅÒẺÊÁÁÙÅ (equivalent discrete time chan nel model) µÒÁÃÙ» ·Õ 4.2 â´Â¢éÍÁÙÅ µèÒ§æ ¨ÐÍÂÙè ã¹â´àÁ¹
D
áÅÐÊÁÁØµÔ ãËé
N (D)
ú¡Ç¹à¡ÒÊì ÊÕ ¢ÒÇẺºÇ¡ (AWGN) ¨Ò¡ÃÙ» ¨Ðä´é ÇèÒ ÊÑÒ³·Õ ǧ¨ÃÀÒ¤ÃѺ ä´é ÃѺ
à» ¹ ÊÑÒ³
P (D)
ÊÒÁÒö
à¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§ÊÁ¡Ò÷ҧ¤³ÔµÈÒʵÃì ¤×Í
P (D) = A(D)C(D) + N (D)
(4.1)
4.2.
ÍÕ¤ÇÍäÅà«ÍÃì
â´Â·Ñ Ç仪èͧÊÑÒ³
67
C(D)
¨ÐÁÕ ÅѡɳÐà» ¹ ǧ¨Ã¡Ãͧ·Õ ÁÕ ¼ÅµÍºÊ¹Í§ÍÔÁ¾ÑÅÊì ¨Ó¡Ñ´ (FIR:
nite impulse response) áÅÐÁÕ ¨Ó¹Ç¹á·ç» ÁÒ¡ (¡èÍ ãËé à¡Ô´ ISI ÁÒ¡) ¶éÒ äÁè ÁÕ ¡ÒÃãªé ÍÕ¤ÇÍäÅà«ÍÃì
F (D)
à¾× ÍÅ´¼Å¡Ãзº¢Í§ ISI ãËé¹éÍÂŧ ǧ¨ÃµÃǨËÒ (detector) ·Õ ãªé¨ÐµéͧÁÕ¤ÇÒÁ«Ñº«é͹ÁÒ¡
à¾× Í ¨Ñ´¡ÒáѺ ISI ¨Ó¹Ç¹ÁÒ¡ à¾ÃÒÐ©Ð¹Ñ ¹ à¾× Í Å´¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒ ¨Ö§ ä´é ÁÕ ¡ÒÃ¹Ó ÍÕ¤ÇÍäÅà«ÍÃì ÁÒãªé §Ò¹ à¾× Í »ÃѺ ÃÙ»ÃèÒ§¢Í§ÊÑÒ³ãËé à» ¹ 仵ÒÁ·ÒÃìà¡çµ ·Õ µéͧ¡Òà (à» ¹ ǧ¨Ã¡Ãͧ Ẻ FIR ·Õ Áըӹǹá·ç»¹éÍÂ) «Ö §¨ÐªèÇÂÅ´¼Å¡Ãзº¢Í§ ISI ãËé¹éÍÂŧä´é ÍÂèÒ§äáçµÒÁ ¡ÒÃ¹Ó ÍÕ¤ÇÍäÅà«ÍÃìÁÒãªé§Ò¹ÁÕ¢éÍàÊÕ ¤×Í (¶éÒÇÒ§ÍÕ¤ÇÍäÅà«ÍÃìäÇéËÅѧǧ¨ÃªÑ¡µÑÇÍÂèÒ§) ¨Ð·ÓãËéà¡Ô´»ÃÔÁÒ³ ˹èǧàÇÅÒ (delay) ¨Ó¹Ç¹ÁÒ¡ã¹ä·ÁÁÔ §ÅÙ» ¡ÅèÒǤ×Í ¨Ó¹Ç¹á·ç»¢Í§ÍÕ¤ÇÍäÅà«ÍÃìÂÔ §ÁÒ¡ »ÃÔÁÒ³ ˹èǧàÇÅÒ¡ç¨ÐÂÔ §ÁÒ¡ «Ö §¨ÐÊ觼ŷÓãËéÍѵÃÒ¡ÒÃÅÙèà¢éÒ (convergence rate) ¢Í§Ãкºä·ÁÁÔ §ÃԤѿàÇÍÃÕ ªéÒŧ ·ÓãËéǧ¨Ãà¿ÊÅçÍ¡ÅÙ» (PLL) äÁèÊÒÁÒöµÔ´µÒÁ¡ÒÃà»ÅÕ Â¹á»Å§à¿ÊáÅФÇÒÁ¶Õ ¢Í§ÊÑÒ³ á͹ÐÅçÍ¡·Õ ¨Ð·Ó¡ÒêѡµÑÇÍÂèÒ§ä´é·Ñ¹ «Ö §ÍÒ¨¨ÐÊ觼ŷÓãËéÕà¡Ô´¡ÒÃÊÙàÊÕ¡Ãкǹ¡ÒÃà¢éҨѧËÇÐä´é
4.2.1
ÍÕ¤ÇÍäÅà«ÍÃìà຺¼ÅµÍºÊ¹Í§àµçÁ
ÍÕ¤ÇÍäÅà«ÍÃìẺ¼ÅµÍºÊ¹Í§àµçÁ (full response equalizer) ËÁÒ¶֧ÍÕ¤ÇÍäÅà«ÍÃì·Õ ¨Ð·ÓãËé¢éÍÁÙÅ àÍÒµì¾Øµ·Õ ä´éÁÕ¤èÒà·èҡѺ ¢éÍÁÙÅÍÔ¹¾Øµ
A(D)
ºÇ¡¡ÑºÊÑҳú¡Ç¹
W (D)
´Ñ§¹Ñ ¹ ¨Ò¡ÃÙ»·Õ 4.2 ¨Ð
ä´é ÇèÒ ÍÕ¤ÇÍäÅà«ÍÃì Ẻ¼ÅµÍºÊ¹Í§àµçÁ ¨ÐÁÕ ¼ÅµÍºÊ¹Í§ÍÔÁ¾ÑÅÊì (impulse response) ã¹â´àÁ¹
D
¤×Í
F (D) = áÅТéÍÁÙÅàÍÒµì¾Øµ
Y (D)
1 C(D)
(4.2)
¢Í§ÍÕ¤ÇÍäÅà«ÍÃì¹Õ ¤×Í
Y (D) = P (D)F (D) á·¹¤èÒ
F (D)
(4.3)
¨Ò¡ÊÁ¡Òà (4.2) ŧã¹ÊÁ¡Òà (4.1) ¨Ðä´é
Y (D) = {A(D)C(D) + N (D)} = A(D) +
N (D) C(D) | {z } W (D)
1 C(D) (4.4)
68
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¹Ñ ¹¤×Í Í§¤ì»ÃСͺ¢Í§ÊÑҳú¡Ç¹·Õ ¨Ðà¢éÒ ä»ã¹Ç§¨ÃµÃǨËÒÊÑÅѡɳì (symbol detector) ¤×Í
W (D) = N (D)/C(D)
¶éÒ ÊÁÁØµÔ ÇèÒ
W (D)
ÁÕ ¤èÒ ¹éÍÂÁÒ¡ ǧ¨ÃµÃǨËÒÊÑÅÑ¡É³ì ·Õ ãªé ¡ç
ÊÒÁÒö໠¹ Ẻ§èÒÂæ ä´é àªè¹ ǧ¨ÃµÃǨËÒ¢Õ´ àÊé¹ áºè§ ẺËÅÒÂÃдѺ (multi level threshold detector) à¾× Í ·Ó¡ÒöʹÃËÑÊ ¢éÍÁÙÅ
Y (D)
ÍÂèÒ§äáçµÒÁ ¢éÍàÊÕ ¢Í§¡ÒÃãªé ÍÕ¤ÇÍäÅà«ÍÃì Ẻ¼Å
µÍºÊ¹Í§àµçÁ ¡ç ¤×Í ÊÑҳú¡Ç¹
W (D)
·Õ ËŧàËÅ×Í ÍÂÙè ÍÒ¨¨Ð¡èÍ ãËé à¡Ô´ »ÃÒ¡®¡ÒÃ³ì ¡ÒâÂÒÂ
ÊÑҳú¡Ç¹ ¹Ñ ¹¤×Í
W (D)
ÁÕ ¤èÒ à» ¹ ¤èÒ Í¹Ñ¹µì ¶éÒ ªèͧÊÑÒ³
C(D)
ÁÕ Ê໡µÃÑÁ ¤èÒ ÈÙ¹Âì
(spectral null) ·Õ ¤ÇÒÁ¶Õ ã´æ à¾ÃÒÐ©Ð¹Ñ ¹ ã¹·Ò§»¯ÔºÑµÔ ¨Ö§äÁè¹ÔÂÁ¹ÓÍÕ¤ÇÍäÅà«ÍÃìẺ¼ÅµÍºÊ¹Í§ àµçÁÁÒãªé§Ò¹ã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì
4.2.2
ÍÕ¤ÇÍäÅà«ÍÃìà຺¼ÅµÍºÊ¹Í§ºÒ§Êèǹ
ÍÕ¤ÇÍäÅà«ÍÃì Ẻ¼ÅµÍºÊ¹Í§ºÒ§Êèǹ (partial response equalizer) ¤×Í ÍÕ¤ÇÍäÅà«ÍÃì ·Õ ÊÒÁÒö à¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§ÊÁ¡Ò÷ҧ¤³ÔµÈÒʵÃìä´é´Ñ§¹Õ
F (D) =
â´Â·Õ
H(D)
H(D) C(D)
(4.5)
¤×Í ¼ÅµÍºÊ¹Í§·ÒÃìà¡çµ (target response) ·Õ µéͧ¡Òà áÅÐàÁ× Íá·¹¤èÒ
F (D)
¹Õ ŧ
ã¹ÊÁ¡Òà (4.3) ¨Ðä´é
H(D) C(D) H(D) = A(D)H(D) + N (D) | {z } C(D) | {z } wanted signal
Y (D) = {A(D)C(D) + N (D)}
(4.6)
W (D)
¹Ñ ¹¤×Í ¢éÍÁÙÅàÍÒµì¾Øµ¢Í§ÍÕ¤ÇÍäÅà«ÍÃìẺ¼ÅµÍºÊ¹Í§ºÒ§Êèǹ¨Ð»ÃСͺ仴éÇ ¢éÍÁÙÅ·Õ µéͧ¡ÒÃ
A(D)H(D)
áÅÐÊÑҳú¡Ç¹
W (D) = N (D)H(D)/C(D)
¨Ò¡ÊÁ¡Òà (4.6) ¨ÐàËç¹ä´éÇèÒ
¢éÍÁÙÅ·Õ µéͧ¡ÒèÐÁÕ ISI ὧÍÂÙè áµèà¹× ͧ¨Ò¡ ǧ¨ÃÀÒ¤ÃѺ·ÃÒºÇèÒ ISI ¹Õ ¤×ÍÍÐäà (à¾ÃÒÐÇèÒà» ¹ ISI ·Õ à¡Ô´¨Ò¡·ÒÃìà¡çµ) ´Ñ§¹Ñ ¹ ISI ¹Õ ÊÒÁÒö·Õ ¨Ð¶Ù¡¨Ñ´¡ÒÃä´é´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ «Ö §¨Ð͸ԺÒµèÍä» ã¹ËÑÇ¢éÍ·Õ 4.3
4.3.
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
69
¹Í¡¨Ò¡¹Õ àÁ× Í ¾Ô¨ÒóÒÊèǹ¢Í§ÊÑҳú¡Ç¹
W (D)
ã¹ÊÁ¡Òà (4.6) ¨Ð¾ºÇèÒ ÊÒà赯 ·Õ
Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì µéͧ¡ÒÃ·Õ ¨Ðä´é ·ÒÃìà¡çµ àªÔ§ ¤ÇÒÁ¶Õ àËÁ×͹¡Ñº ¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§ªèͧÊÑÒ³ ·ÓãËé
W (D)
ÁÕ ÅѡɳÐà» ¹ ÊÑҳú¡Ç¹à¡ÒÊì ÊÕ ¢ÒÇ
H(D) = C(D)
áÅéÇ ¨Ðä´é ÇèÒ
W (D) = N (D)
C(D)
N (D)
H(D)
·Õ ÁÕ ¼ÅµÍºÊ¹Í§
ãËé ÁÒ¡·Õ ÊØ´ ¡ç à¾× Í ·Õ ÇèÒ¨Ðä´é
ãËé ÁÒ¡·Õ ÊØ´ ·Ñ §¹Õ à» ¹ à¾ÃÒÐÇèÒ ¶éÒ
«Ö § ¶×Í ÇèÒ à» ¹ à§× ͹ä¢ËÅÑ¡ ·Õ ¨Ð·ÓãËé ǧ¨ÃµÃǨËÒ
ÇÕà·ÍÃìºÔÊÒÁÒö·Ó§Ò¹ä´éÍÂèÒ§ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡·Õ ÊØ´ ËÃ×Í¡ÅèÒÇÍÕ¡¹ÑÂË¹Ö §¤×Í Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨Ð¶Ù¡¾Ô¨ÒóÒÇèÒà» ¹ ǧ¨ÃµÃǨËÒ·Õ àËÁÒÐ·Õ ÊØ´ (opimal detector) ¶éÒͧ¤ì»ÃСͺ¢Í§ÊÑÒ³ ú¡Ç¹·Õ ´éÒ¹¢Òà¢éҢͧǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔà» ¹ÊÑҳú¡Ç¹à¡ÒÊìÊÕ¢ÒÇ [15] ´Ñ§¹Ñ ¹ ÍÒ¨¨ÐÊÃØ» ä´éÇèÒ ¶éҼŵͺʹͧàªÔ§¤ÇÒÁ¶Õ ¢Í§·ÒÃìà¡çµàËÁ×͹¡Ñº¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ ¢Í§ªèͧÊÑÒ³ÁÒ¡ à·èÒã´ »ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкºã¹ÃÙ»¢Í§ÍѵÃÒ¢éͼԴ¾ÅÒ´ºÔµ (BER) ÇÑ´·Õ ´éÒ¹¢ÒÍÍ¡¢Í§Ç§¨ÃµÃǨËÒ ÇÕà·ÍÃìºÔ¡ç¨Ð´ÕÁÒ¡¢Ö ¹à·èÒ¹Ñ ¹
4.3
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¤×Í Ç§¨ÃµÃǨËÒÅӴѺ (sequence detector) ·Õ ÊÃéÒ§â´Âãªé ÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ (Viterbi algorithm) [15] à¾× Íãªé㹡ÒöʹÃËÑÊ¢éÍÁÙÅ·Õ ¶Ù¡à¢éÒÃËÑÊ´éÇ ÃËÑʤ͹âÇÅ٪ѹ (convo lutional code) [7] à·èÒ¹Ñ ¹ ã¹·Ò§»¯ÔºÑµÔáÅéÇ ªèͧÊÑÒ³ÊÒÁÒö·Õ ¨Ð¶Ù¡¾Ô¨ÒóÒÇèÒà» ¹ÃËÑʤ͹ âÇÅ٪ѹ»ÃÐàÀ·Ë¹Ö §·Õ ÁÕÍѵÃÒÃËÑÊ (code rate) à·èҡѺ¤èÒË¹Ö § (¹Ñ ¹¤×Í ¢éÍÁÙÅÍÔ¹¾Øµ 1 ºÔµ àÁ× Íà¢éÒÃËÑÊ áÅéǨÐä´é¢éÍÁÙÅàÍÒµì¾ØµÍÍ¡ÁÒ 1 ºÔµàªè¹¡Ñ¹) ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔÁÕ¤ÇÒÁÊÒÁÒö·Õ ¨Ñ´¡ÒáѺ ISI ·Õ ὧÍÂÙèã¹¢éÍÁÙÅ·Õ ¨Ð·Ó¡ÒöʹÃËÑÊä´éÍÂèÒ§ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ â´Â·Õ ¶éÒ ISI ÂÔ §ÁÒ¡ ¤ÇÒÁ«Ñº«é͹¢Í§ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¡ç¨ÐÂÔ §ÁÒ¡ áÅжéÒ ISI ¹éÍ ¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¡ç¨Ð¹éÍ à¹× ͧ¨Ò¡ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÁÕ ¤ÇÒÁ«Ñº«é͹ÁÒ¡¡ÇèÒ Ç§¨ÃµÃǨËÒẺ§èÒ (simple detector) àªè¹ ǧ¨ÃµÃǨËÒ¢Õ´ àÊé¹ áºè§ ẺËÅÒÂÃдѺ 㹡ÒÃ·Õ ¨ÐµÑ´ÊÔ¹ã¨ÇèÒ ¨Ð¹Óǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÁÒ ãªé §Ò¹ã¹ÃкºËÃ×Í äÁè ¹Ñ ¹ ãËé ¾Ô¨ÒóҨҡÃÙ» ·Õ 4.3 ´Ñ§µèÍä»¹Õ ¨Ò¡ÃÙ» ·Õ 4.3(a) ¶éÒ ªèͧÊÑÒ³äÁè ÁÕ ISI ǧ¨ÃÀÒ¤ÃѺ ¡ç ÊÒÁÒö¹Óǧ¨ÃµÃǨËÒẺ§èÒÂÁÒãªé §Ò¹ä´é àÅ ¶éÒ ªèͧÊÑÒ³ÁÕ ISI ¹éÍ µÒÁÃÙ» ·Õ 4.3(b) ǧ¨ÃÀÒ¤ÃѺ ¡ç ÊÒÁÒö¹Óǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÁÒãªé §Ò¹ä´é àÅ áµè ¶éÒ ªèͧÊÑÒ³ ÁÕ ISI ¨Ó¹Ç¹ÁÒ¡µÒÁÃÙ» ·Õ 4.3(c) ǧ¨ÃÀÒ¤ÃѺ ¡ç ¤ÇÃ·Õ ¹ÓÍÕ¤ÇÍäÅà«ÍÃì ÁÒãªé §Ò¹ à¾× Í Å´¼Å¡Ãзº
70
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
AWGN
(a)
ak
simple detector
channel no ISI
âk
AWGN
(b)
ak
channel with low ISI C(D)
Viterbi detector C(D)
âk
AWGN
ak (c)
channel with severe ISI C(D)
equalizer
Viterbi detector H(D)
âk
target H(D) ÃÙ»·Õ 4.3: µÑÇÍÂèҧẺ¨ÓÅͧªèͧÊÑÒ³·Õ äÁèµèÍà¹× ͧ·Ò§àÇÅÒẺÊÁÁÙÅÅѡɳеèÒ§æ
¢Í§ ISI ãËé ¹éÍÂŧ ¨Ò¡¹Ñ ¹ ¨Ö§ ¤èÍÂÊè§ ¢éÍÁÙÅ àÍÒµì¾Øµ ·Õ ä´é ¨Ò¡ÍÕ¤ÇÍäÅà«ÍÃì ä»·Ó¡ÒöʹÃËÑÊ ¢éÍÁÙÅ ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¶×ÍÇèÒà» ¹Ç§¨ÃµÃǨËÒ¢éÍÁÙÅ·Õ ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ áÅж١¹ÓÁÒãªé§Ò¹ã¹ËÅÒÂæ §Ò¹»ÃÐÂØ¡µì ÃÇÁ·Ñ §ã¹Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì â´Â·Õ ËÅÑ¡¡Ò÷ӧҹ¢Í§ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨ÐÍÂÙè º¹¾× ¹°Ò¹¢Í§ á¼¹ÀÒ¾à·ÃÅÅÔÊ à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´
(trellis diagram) «Ö § ÊÃéÒ§ÁÒ¨Ò¡
(FSM: nite state machine) ´Ñ§¹Ñ ¹ ¡è͹·Õ ¨Ð͸ԺÒÂËÅÑ¡¡Ò÷ӧҹ¢Í§
ÍÑÅ¡ÍÃÔ·ÖÁ ÇÕà·ÍÃìºÔ ¼ÙéÍèÒ¹¤ÇèзӤÇÒÁà¢éÒã¨à¡Õ ÂǡѺ ÇÔ¸Õ¡ÒÃÊÃéÒ§à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ áÅÐá¼¹ÀÒ¾ à·ÃÅÅÔÊ¡è͹ «Ö §ÁÕÃÒÂÅÐàÍÕ´´Ñ§¹Õ
4.3.
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
71
rk
H(D)
ak
D
{0,1} ÃÙ»·Õ 4.4: á¼¹ÀÒ¾ªèͧÊÑҳẺ PR1,
µÒÃÒ§·Õ 4.1: ¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§
4.3.1
ak
áÅÐ
rk
H(D) = 1 + D
¢Í§ªèͧÊÑÒ³
ak
ak−1
rk
0
0
0
0
1
1
1
0
1
1
1
2
H(D) = 1 + D
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´
¾Ô¨ÒóҪèͧÊÑÒ³
H(D) = 1 + D
ã¹ÃÙ»·Õ 4.4 àÁ× Í ¢éÍÁÙźԵÍÔ¹¾Øµ
ak ∈ {0, 1}
à¢éÒä»ã¹ªèͧÊÑÒ³·ÓãËéä´éà» ¹¢éÍÁÙÅàÍÒµì¾ØµªèͧÊÑÒ³ (channel output) à¢Õ¹µÒÃÒ§áÊ´§¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§¢éÍÁÙÅÍÔ¹¾Øµ â´Â·Õ ¢éÍÁÙźԵ
ak−1
register) 㹺ÅçÍ¡
ak
áÅТéÍÁÙÅàÍÒµì¾Øµ
rk
¶Ù¡Ê觼èÒ¹
rk ∈ {0, 1, 2}
¶éÒ
¨Ðä´éµÒÁµÒÃÒ§·Õ 4.1
ÍÒ¨¨Ð¶Ù¡¾Ô¨ÒóÒÇèÒà» ¹¢éÍÁÙÅ·Õ ËŧàËÅ×ÍÍÂÙèã¹ àèÔÊàµÍÃìẺàÅ× Í¹ (shift
D
¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§
ak
áÅÐ
rk
ã¹µÒÃÒ§·Õ 4.1 ÊÒÁÒöáÊ´§ãËéÍÂÙèã¹ÃÙ»¢Í§ à¤Ã× Í§Ê¶Ò¹Ð
¨Ó¡Ñ´ (FSM: nite state machine) «Ö § ¡ç ¤×Í áºº¨ÓÅͧ¢Í§¡Òäӹdz·Õ áÊ´§ãËé àËç¹ ¶Ö§ ¡Òà à»ÅÕ Â¹á»Å§¢Í§ ¢éÍÁÙÅÍÔ¹¾Øµ, ʶҹÐàÃÔ Áµé¹ (start state), ʶҹеèÍä» (next state), áÅТéÍÁÙÅ àÍÒµì¾ØµªèͧÊÑÒ³ ´Ñ§áÊ´§ã¹ÃÙ»·Õ 4.5 â´Â·Õ ¢éÍÁÙźԵ 0 áÅкԵ 1 ·Õ ÍÂÙèã¹Ç§¡ÅÁ ¤×Í ¤èÒ·Õ ÍÂÙè
72
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
1/1 X/Y is denoted as:
1
0
0/0
1/2
ak Y = the output bit rk X = the input bit
0/1 ÃÙ»·Õ 4.5: à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ªèͧÊÑÒ³ PR1,
ã¹àèÔÊàµÍÃìẺàÅ× Í¹ ³ àÇÅÒ¢³Ð¹Ñ ¹, ªèͧÊÑÒ³
rk
X
¤×Í ¤èÒ¢éÍÁÙÅÍÔ¹¾Øµ
H(D) = 1 + D
ak ,
áÅÐ
Y
¤×Í ¤èÒ¢éÍÁÙÅàÍÒµì¾Øµ
µÑÇÍÂèÒ§àªè¹ ÊÁÁصÔÇèÒ ³ µÍ¹¹Õ ¤èÒ·Õ ÍÂÙèã¹àèÔÊàµÍÃìẺàÅ× Í¹ ¤×Í ¤èÒ 1 ¶éÒ¢éÍÁÙÅ
ÍÔ¹¾Øµ ·Õ à¢éÒ ÁÒ㹪èͧÊÑÒ³ ¤×Í
ak = 0
¨Ðä´é ÇèÒ ¤èÒ ·Õ ÍÂÙè ã¹àèÔÊàµÍÃì ẺàÅ× Í¹àÁ× Í àÇÅÒ¼èÒ¹ä»
Ë¹Ö §Ë¹èÇ (ËÃ×Í Ê¶Ò¹ÐµèÍä») ¤×Í ¤èÒ 0 â´Â¨Ðä´é¢éÍÁÙÅàÍÒµì¾ØµªèͧÊÑÒ³
rk = 1
㹷ӹͧ
à´ÕÂǡѹ ¶éҵ͹¹Õ ¤èÒ·Õ ÍÂÙèã¹àèÔÊàµÍÃìẺàÅ× Í¹ ¤×Í ¤èÒ 0 áÅТéÍÁÙÅÍÔ¹¾Øµ·Õ à¢éÒÁÒ㹪èͧÊÑÒ³ ¤×Í
ak = 0
¨Ðä´é ÇèÒ ¤èÒ ·Õ ÍÂÙè ã¹àèÔÊàµÍÃì ẺàÅ× Í¹àÁ× Í àÇÅÒ¼èÒ¹ä»Ë¹Ö § ˹èÇ ¤×Í ¤èÒ 0 â´Â¨Ðä´é
rk = 0
¢éÍÁÙÅàÍÒµì¾ØµªèͧÊÑÒ³
ËÁÒÂà˵Ø
¶éÒ ¡Ó˹´ãËé
|A|
à» ¹µé¹
¤×Í ¨Ó¹Ç¹¤èÒ ·Õ à» ¹ ä»ä´é ·Ñ §ËÁ´¢Í§¢éÍÁÙÅ ÍÔ¹¾Øµ
ak ,
áÅÐ
ν
˹èǤÇÒÁ¨Ó¢Í§ªèͧÊÑÒ³ ´Ñ§¹Ñ ¹ ¨Ó¹Ç¹Ê¶Ò¹Ð (state) ·Ñ §ËÁ´ã¹à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ ¤×Í àªè¹ Ẻ¨ÓÅͧªèͧÊÑÒ³ã¹ÃÙ»·Õ 4.4 ¨Ðä´éÇèÒ
|A| = 2
¨Ó¡Ñ´¢Í§ªèͧÊÑÒ³ PR1 ¨ÐÁÕ¨Ó¹Ç¹Ê¶Ò¹Ð·Ñ §ËÁ´
4.3.2
áÅÐ
21 = 2
ν=1
¤×Í
|A|ν
à¾ÃÒÐ©Ð¹Ñ ¹ à¤Ã× Í§Ê¶Ò¹Ð
ʶҹÐ
à༹ÀÒ¾à·ÃÅÅÔÊ
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´·Õ áÊ´§ã¹ÃÙ»·Õ 4.5 ÊÒÁÒöáÊ´§ãËéÍÂÙèã¹ÃÙ»¢Í§á¼¹ÀÒ¾à·ÃÅÅÔÊä´é ´Ñ§áÊ´§ã¹ ÃÙ»·Õ 4.6 â´Â·Õ ¨Ø´µèÍ (node) ·Ò§´éÒ¹«éÒÂÁ×Í ¤×Í Ê¶Ò¹ÐàÃÔ Áµé¹ («Ö §ÁÕ¤èÒà·èҡѺ ´éÒ¹¢ÇÒÁ×Í ¤×Í Ê¶Ò¹ÐµèÍä», ÅÙ¡ÈÃàÊé¹»Ðãªéá·¹¡ÒÃÊ觢éÍÁÙÅÍÔ¹¾Øµ ¡ÒÃÊ觢éÍÁÙÅÍÔ¹¾Øµ
ak = 1,
ak = 0,
ak−1 ),
¨Ø´µèÍ·Ò§
ÅÙ¡ÈÃàÊé¹·Öºãªéá·¹
áÅеÑÇàÅ¢·Õ áÊ´§ÍÂÙèã¹áµèÅÐàÊé¹ÊÒ¢Ò (branch) ¤×Í ¤èÒ¢éÍÁÙÅàÍÒµì¾Øµ
4.3.
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
73
[ ak −1]
0
0
0
1
ak = 0
1
1
ak = 1
1
2
ÃÙ»·Õ 4.6: á¼¹ÀÒ¾à·ÃÅÅÔʢͧªèͧÊÑÒ³ PR1,
ªèͧÊÑÒ³
rk
µÑÇÍÂèÒ§àªè¹ ¶éÒʶҹÐàÃÔ Áµé¹ ¤×ͤèÒ 0 àÁ× Í¢éÍÁÙÅÍÔ¹¾Øµ
·ÓãËéä´é¢éÍÁÙÅàÍÒµì¾ØµªèͧÊÑÒ³
rk = 1
àÃÔ Áµé¹ ¤×Í ¤èÒ 1 àÁ× Í ¢éÍÁÙÅ ÍÔ¹¾Øµ
ak = 1
rk = 2
H(D) = 1 + D
ak = 1
à¢éÒÁÒã¹Ãкº ¨Ð
áÅÐʶҹеèÍä» ¤×ͤèÒ 1 㹷ӹͧà´ÕÂǡѹ ¶éÒʶҹРà¢éÒ ÁÒã¹Ãкº ¨Ð·ÓãËé ä´é ¢éÍÁÙÅ àÍÒµì¾Øµ ªèͧÊÑÒ³
áÅÐʶҹеèÍä» ¤×ͤèÒ 1
㹷ӹͧà´ÕÂǡѹ ÃÙ»·Õ 4.7 áÊ´§µÑÇÍÂèҧἹÀÒ¾à·ÃÅÅÔʢͧªèͧÊÑÒ³ EPR4 (extended PR4),
H(D) = 1 + D − D2 − D3 ,
àÁ× Í ¢éÍÁÙÅÍÔ¹¾Øµ
¢Í§á¼¹ÀÒ¾à·ÃÅÅÔʨÐÁըӹǹà·èҡѺ
{ak } ∈ {0, 1}
|A|ν = 23 = 8
«Ö §ã¹¡Ã³Õ¹Õ ¨Ó¹Ç¹Ê¶Ò¹Ð
ʶҹРà¹× ͧ¨Ò¡ ªèͧÊÑÒ³ EPR4 ÁÕ
¨Ó¹Ç¹Ë¹èǤÇÒÁ¨Óà·èҡѺ 3 ˹èÇÂ
µÑÇÍÂèÒ§·Õ 4.1
¡Ó˹´ãËé¢éÍÁÙÅÍÔ¹¾Øµ
H(D) = 1 − D2 ,
ak ∈ {−1, 1}
áÅéÇä´é¢éÍÁÙÅàÍÒµì¾Øµà» ¹
¶Ù¡Ê觼èÒ¹à¢éÒä»ÂѧªèͧÊÑҳẺ PR4,
rk ∈ {−2, 0, 2}
¨§
¡) ÇÒ´á¼¹ÀÒ¾ºÅçÍ¡¢Í§ªèͧÊÑҳẺ PR4 ¹Õ (¤ÅéÒ¡ѺÃÙ»·Õ 4.4)
¢) à¢Õ¹ÊÁ¡ÒÃáÊ´§¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§¢éÍÁÙÅÍÔ¹¾ØµáÅТéÍÁÙÅàÍÒµì¾Øµ¢Í§ªèͧÊÑÒ³¹Õ
¤) ÊÃéÒ§à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´
§) ÊÃéҧἹÀÒ¾à·ÃÅÅÔÊ
74
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
[ak −1 ak −2 ak −3] (0)
0
000
1
(1)
100
(2)
010
(3)
110
(4)
001
(5)
101
1 2 -1 0 1 -1
0 0 0 1 -2
(6)
011
(7)
111
-1 -1
ak = 1 ak = 0
0
yk ÃÙ»·Õ 4.7: á¼¹ÀÒ¾à·ÃÅÅÔʢͧªèͧÊÑÒ³ EPR4,
H(D) = 1 + D − D2 − D3
ÇÔ¸Õ·Ó ¡)
¨Ò¡ªèͧÊÑÒ³·Õ ¡Ó˹´ãËé ÊÒÁÒöáÊ´§à» ¹á¼¹ÀÒ¾ºÅçÍ¡¢Í§ªèͧÊÑҳẺ PR4 ä´é
´Ñ§ÃÙ»·Õ 4.8(a) ¢)
à¹× ͧ¨Ò¡
H(D) =
P k
hk Dk = 1 − D2
´Ñ§¹Ñ ¹ ÊÁ¡ÒÃáÊ´§¤ÇÒÁÊÑÁ¾Ñ¹¸ì ÃÐËÇèÒ§¢éÍÁÙÅ
ÍÔ¹¾ØµáÅТéÍÁÙÅàÍÒµì¾Øµ¢Í§ªèͧÊÑÒ³¹Õ ¤×Í
rk = ak ∗ hk = ak − ak−2 ¤)
¨Ò¡ÊÁ¡ÒÃáÊ´§¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§¢éÍÁÙÅÍÔ¹¾ØµáÅТéÍÁÙÅàÍÒµì¾Øµ·Õ ä´éã¹¢éÍ ¢) à¤Ã× Í§Ê¶Ò¹Ð
¨Ó¡Ñ´ÊÒÁÒöáÊ´§ä´éµÒÁÃÙ»·Õ 4.8(b)
4.3.
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
75
[ ak −1ak −2 ]
rk ak
D
{±1} (a) Block diagram: PR4
1 -11
1/2 -1/0
-1 - 1-1 -
1/0
-1/0
0
-1 -1
D
2
1 -1
1/2
-1 1 11
0 -2 0
1/0
-1/-2
-1/-2
0 2 -2
1 1
ak = -1 ak = 1
-1111
(b) FSM
(c) Trellis diagram
ÃÙ»·Õ 4.8: (a) á¼¹ÀÒ¾ºÅçÍ¡, (b) à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´, áÅÐ (c) á¼¹ÀÒ¾à·ÃÅÅÔÊ ¢Í§ªèͧÊÑÒ³
H(D) = 1 − D2
§)
á¼¹ÀÒ¾à·ÃÅÅÔÊÊÒÁÒöÊÃéÒ§ä´é¨Ò¡à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ ´Ñ§áÊ´§ã¹ÃÙ»·Õ 4.8(c)
4.3.3
ÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ
¨Ò¡à«µ¢Í§ªØ´µÑÇÍÑ¡ÉÃ
H(D) µÒÁÃÙ»·Õ 4.9 â´Â·Õ ak ¤×Í ¢éÍÁÙźԵÍÔ¹¾Øµ·Õ ¶Ù¡àÅ×Í¡ÁÒ P A, H(D) = νk=0 hk Dk ¤×Í ªèͧÊÑÒ³ÇÔÂص (discrete channel), hk
¤×Í ¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô µÑÇ ·Õ
k, ν
¾Ô¨ÒóÒẺ¨ÓÅͧªèͧÊÑÒ³
¤×Í Ë¹èǤÇÒÁ¨Ó¢Í§ªèͧÊÑÒ³,
nk
¤×Í ÊÑҳú¡Ç¹à¡ÒÊì ÊÕ
¢ÒÇẺºÇ¡ (AWGN) ·Õ ÁÕ¤èÒà©ÅÕ Âà·èҡѺ¤èÒÈÙ¹Âì áÅФÇÒÁá»Ã»Ãǹà·èҡѺ¤èÒ ä´é´éÇÂ
nk ∼ N (0, σ 2 ), rk
ÍÔ¹¾Øµ
{ak }
(â´Â·Ñ Çä»
¤×Í ¢éÍÁÙÅàÍÒµì¾ØµªèͧÊÑÒ³, áÅÐ
L = 4096
L
σ2
ËÃ×Íà¢Õ¹᷹
¤×Í ¤ÇÒÁÂÒǢͧÅӴѺ¢éÍÁÙÅ
ºÔµ ÊÓËÃѺ ¢éÍÁÙÅ 1 à«¡àµÍÃì), áÅÐ
âk
¤×Í ¤èÒ »ÃÐÁÒ³¢Í§
76
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
nk ak
rk
H(D)
yk
ÃÙ»·Õ 4.9: Ẻ¨ÓÅͧªèͧÊÑÒ³
Viterbi detector
H(D)
time k
âk
¾ÃéÍÁǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
k+1
u
(u, q) q k-th stage
ÃÙ»·Õ 4.10: ¤Ó͸ԺÒÂá¼¹ÀÒ¾à·ÃÅÅÔÊ
¢éÍÁÙźԵÍÔ¹¾Øµ
ak
·Õ ä´é¨Ò¡¡ÒöʹÃËÑÊ¢éÍÁÙÅâ´Âǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
¡è͹·Õ ¨Ð͸ԺÒÂËÅÑ¡¡Ò÷ӧҹ¢Í§ÍÑÅ¡ÍÃÔ·ÖÁ ÇÕà·ÍÃìºÔ ¨Ð¹ÔÂÒÁÊÑÅÑ¡É³ì µÒÁÃÙ» ·Õ 4.10 ´Ñ§¹Õ ãËé
Ψk = [ak ak−1 . . . ak−ν+1 ]
³ àÇÅÒ
k ), Q = |A|ν
¢Í§¢éÍÁÙÅ ÍÔ¹¾Øµ,
ν
¤×Í Ê¶Ò¹Ð ³ àÇÅÒ
u
(ËÃ×ͤèÒ·Õ ÍÂÙèã¹àèÔÊàµÍÃìẺàÅ× Í¹·Ñ §ËÁ´
¤×Í ¨Ó¹Ç¹Ê¶Ò¹Ð·Ñ §ËÁ´·Õ à» ¹ä»ä´é,
|A|
¤×Í Ë¹èǤÇÒÁ¨Ó¢Í§ªèͧÊÑÒ³, ÃÐÂзÕ
ÊÒ¢Ò·Õ à» ¹ ä»ä´é ·Ñ §ËÁ´ ³ àÇÅÒ·Õ Ê¶Ò¹Ð
k
ä»ÂѧʶҹÐ
k,
áÅÐ
(u, q)
k (k th
stage) ¤×Í ¡ÅØèÁ ¢Í§àÊé¹
¤×Í ÊÑÅÑ¡É³ì ·Õ ãªé á·¹¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð¨Ò¡
q
¾Ô¨ÒóÒá¼¹ÀÒ¾à·ÃÅÅÔʢͧªèͧÊÑÒ³ PR4 ¹Ñ ¹¤×Í ÁÕ¨Ó¹Ç¹Ê¶Ò¹Ð·Ñ §ËÁ´
¤×Í ¨Ó¹Ç¹¤èÒ·Õ à» ¹ä»ä´é·Ñ §ËÁ´
Q = 22 = 4
H(D) = 1 − D2
µÒÁÃÙ»·Õ 4.11 «Ö §
ʶҹР·Õ áÊ´§´éÇÂÊÑÅѡɳì (0), (1), (2), áÅÐ (3) àÁ× Í
4.3.
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
77
time k (0) -1 -1
k+1 0 2
(1) 1 -1
0 2 -2
(2) -1 1
λk (1, 2)
Φ + (2)
0
k 1
π k +1 (2)
-2
(3) 1 1
0
ak = 1 ak = -1
yk ÃÙ»·Õ 4.11: á¼¹ÀÒ¾à·ÃÅÅÔʢͧªèͧÊÑÒ³ PR4,
¢éÍÁÙÅ ÍÔ¹¾Øµ
ak ∈ {−1, 1}
㹡Ò÷ӧҹ¢Í§ÍÑÅ¡ÍÃÔ·ÖÁ ÇÕà·ÍÃìºÔ ÊÔ § ·Õ µéͧ¤Ó¹Ç³·Ø¡ ªèǧàÇÅÒ ¤×Í
¤èÒ àÁµÃÔ¡ ÊÒ¢Ò (branch metric) ³ àÇÅÒ
λk (u, q),
k
¢Í§¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð¨Ò¡Ê¶Ò¹Ð
k+1
¤èÒ àÁµÃÔ¡ àÊé¹·Ò§ (path metric) ³ àÇÅÒ
(predecessor) ÊÓËÃѺʶҹÐ
H(D) = 1 − D2
q
³ àÇÅÒ
k + 1, πk+1 (q),
·Õ ʶҹÐ
u
ä»Âѧ ʶҹÐ
q , Φk+1 (q),
áÅеÑǹÓ˹éÒ
«Ö §¨Ðà¡çº¤èÒʶҹÐàÃÔ Áµé¹·Õ à» ¹¼Å·ÓãËé
à¡Ô´àÊé¹·Ò§¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð·Õ ´Õ·Õ ÊØ´ (best transition) àªè¹ ¾Ô¨ÒÃ³Ò·Õ Ê¶Ò¹Ð (2) ³ àÇÅÒ ¨ÐÁÕàÊé¹·Ò§¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð 2 àÊé¹·Ò§ ¤×Í
(1, 2)
áÅÐ
(3, 2)
¡ÒÃàÅ×Í¡àÊé¹·Ò§à¾Õ§àÊé¹·Ò§à´ÕÂÇ·Õ ÁҶ֧ʶҹР(2) ³ àÇÅÒ àÊé¹·Ò§¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð·Õ ´Õ·Õ ÊØ´ ¨Ðä´éÇèÒ
πk+1 (2)
q,
k+1
ÊÁÁصÔÇèÒ ÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ¨Ð·Ó
k+1
ÊÁÁصÔÇèÒ àÊé¹·Ò§
(1, 2)
¤×Í
= 1 ¹Ñ ¹àͧ
à» ¹ ·Õ ·ÃÒº¡Ñ¹ ÇèÒ Ç§¨ÃµÃǨËÒ·Õ ·ÓãËé ¤ÇÒÁ¹èÒ¨Ðà» ¹ ¢Í§¢éͼԴ¾ÅÒ´¢Í§ÅӴѺ ¢éÍÁÙÅ ·Ñ § ÅӴѺ ÁÕ ¤èÒ¹éÍÂ·Õ ÊØ´ ¤×Í Ç§¨ÃµÃǨËÒÅӴѺ·Õ ¤ÇèÐà» ¹ÁÒ¡ÊØ´ (MLSD: maximum likelihood sequence detector) «Ö §ÊÒÁÒöÊÃéÒ§ä´éâ´ÂãªéÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ ¨Ò¡áºº¨ÓÅͧªèͧÊÑÒ³ã¹ÃÙ»·Õ 4.9 ǧ¨Ã µÃǨËÒÇÕà·ÍÃìºÔ ¨ÐàÅ×Í¡ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ
{ak }
·Õ ·ÓãËé ¤ÇÒÁ¹èÒ¨Ðà» ¹ ¢Í§ÅӴѺ ¢éÍÁÙÅ
{yk }
àÁ× Í
78
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¡Ó˹´ÅӴѺ¢éÍÁÙÅ
{ak }
ÁÒãËé ¹Ñ ¹¤×Í
(
L+ν−1 1 X exp − 2 |yk − rk |2 2σ
1
p(y|a) = ³√ ´L+ν 2πσ 2
a
(4.7)
k=0
ÁÕ¤èÒÁÒ¡·Õ ÊØ´ ÊÁ¡Òà (4.7) ä´éÁÒ¨Ò¡¤ÇÒÁ¨ÃÔ§·Õ ÇèÒ ÊÑҳú¡Ç¹ ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ
)
nk ∼ N (0, σ 2 ) áÅÐàÁ× Í¡Ó˹´
ÁÒãËé áÊ´§ÇèÒ Ãкº·ÃÒºÇèÒ ÅӴѺ ¢éÍÁÙÅ àÍÒµì¾Øµ
r
¤×Í ÍÐäà ´Ñ§¹Ñ ¹ ¢éÍÁÙÅ
yk
¨ÐÁÕ¿ §¡ìªÑ¹¤ÇÒÁ˹Òá¹è¹¤ÇÒÁ¹èÒ¨Ðà» ¹áººà¡ÒÊìà«Õ¹ (Gaussian probability density function) àËÁ×͹¡Ñº
nk
·Õ ÁÕ¤èÒà©ÅÕ Âà·èҡѺ¤èÒ
rk
áÅФèÒ¤ÇÒÁá»Ã»Ãǹà·èҡѺ
σ2
àÁ× ÍãÊèÅÍ¡ÒÃÔ·ÖÁ¸ÃÃÁªÒµÔ (natural logarithm) ·Ñ §Êͧ¢éÒ§¢Í§ÊÁ¡Òà (4.7) ¨Ðä´éà» ¹
ln {p(y|a)} = ln
1
³√ ´L+ν 2πσ 2
−
L+ν−1 1 X |yk − rk |2 2σ 2
(4.8)
k=0
Êѧࡵ¨Ð¾ºÇèÒ ¡Ò÷ÓãËé ÊÁ¡Òà (4.8) ÁÕ ¤èÒ ÁÒ¡·Õ ÊØ´ ÁÕ ¼Åà·Õºà·èÒ ¡Ñº ¡Ò÷ÓãËé ¾¨¹ì ·Õ Êͧ·Ò§´éÒ¹ ¢ÇÒÁ×Í ¢Í§ÊÁ¡Òà (4.8) ÁÕ ¤èÒ ¹éÍÂ·Õ ÊØ´ à¹× ͧ¨Ò¡ ¾¨¹ì ·Õ Ë¹Ö § à»ÃÕºàÊÁ×͹¡Ñº ¤èÒ¤§·Õ à¾ÃÒÐ©Ð¹Ñ ¹ ¡Ò÷ÓãËéÊÁ¡Òà (4.8) ÁÕ¤èÒÁÒ¡·Õ ÊØ´¨ÐÁÕ¤èÒà·èҡѺ¡Ò÷ÓãËéàÁµÃÔ¡ (metric)
L+ν−1 X
|yk − rk |2
(4.9)
k=0
ÁÕ¤èÒ¹éÍÂ·Õ ÊØ´ ´Ñ§¹Ñ ¹ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¨ÐàÅ×Í¡ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ·Õ ·ÓãËéÊÁ¡Òà (4.9) ÁÕ¤èÒ¹éÍ ·Õ ÊØ´ Êѧࡵ¨Ò¡ÊÁ¡Òà (4.9) ¨Ð¾ºÇèÒ ¾¨¹ì
|yk − rk |2
1
¡ç ¤×Í ¤èÒ ÃÐÂзҧ¡ÓÅѧ Êͧà©ÅÕ Â
(MSD:
mean squared distance) [10] àÁµÃÔ¡ã¹ÊÁ¡Òà (4.9) ÊÒÁÒö·ÓãËéÁÕ¤èÒ¹éÍÂ·Õ ÊØ´ä´é â´Â¡Òäé¹ËÒ àÊé¹·Ò§ (path) ·Õ ÁÕ ¤èÒ àÁµÃÔ¡ ¹éÍÂ·Õ ÊØ´ µÒÁá¼¹ÀÒ¾à·ÃÅÅÔÊ àÁ× Í àÁµÃÔ¡ àÊé¹·Ò§ÁÕ ¤èÒ à·èÒ ¡Ñº ¼ÅÃÇÁ ¢Í§àÁµÃÔ¡ÊÒ¢Ò â´Â·Õ àÁµÃÔ¡ÊҢҢͧ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð¨Ò¡Ê¶Ò¹Ð
λk (u, q) = |yk − r̂k (u, q)|2 1
ÈÖ¡ÉÒÃÒÂÅÐàÍÕ´à¾Ô ÁàµÔÁä´éã¹ËÑÇ¢éÍ·Õ 6.6.11 ã¹ [10]
u
ä»ÂѧʶҹÐ
q
¨Ð¹ÔÂÒÁâ´Â
(4.10)
4.3.
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
79
(A 1) ¡Ó˹´¤èÒàÃÔ Áµé¹¢Í§àÁµÃÔ¡àÊé¹·Ò§ (A 2) For
k = 0, 1, . . . , L + ν − 1
(A 3)
For
Φ0 (p) = 0
ÊÓËÃѺ·Ø¡¤èÒ
q = 0, 1, . . . , Q − 1
(A 4)
λk (p, q) = |yk − r̂(p, q)|2
(A 5)
πk+1 (q) = arg minp {Φk (p) + λk (p, q)}
(A 6)
Φk+1 (q) = Φk (πk+1 (q)) + λk (πk+1 (q), q)
(A 7)
Sk+1 (q) = [Sk (πk+1 (q)) | πk+1 (q)]
(A 8)
p
for
∀p
End
(A 9) End (A 10) ¶Í´ÃËÑÊ¢éÍÁÙÅÍÔ¹¾Øµ
â
¨Ò¡àÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè·Õ ÁÕ¤èÒ
ΦL+ν
¹éÍÂ·Õ ÊØ´
ÃÙ»·Õ 4.12: ¢Ñ ¹µÍ¹¡Ò÷ӧҹ¢Í§ÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ
àÁ× Í
r̂k (u, q)
¤×Í ¢éÍÁÙÅàÍÒµì¾ØµªèͧÊÑÒ³·Õ ÊÍ´¤Åéͧ¡Ñº
(u, q)
áÅÐàÁµÃÔ¡àÊé¹·Ò§ÊÒÁÒöËÒä´é
¨Ò¡
Φk+1 (q) =
k X
λi
(4.11)
i=0 ÃÙ» ·Õ 4.12 áÊ´§¢Ñ ¹µÍ¹¡Ò÷ӧҹ¢Í§ÍÑÅ¡ÍÃÔ·ÖÁ ÇÕà·ÍÃìºÔ ¢Í§ªèͧÊÑÒ³
H(D) = 1 − D2
µÑÇÍÂèÒ§àªè¹ ¨Ò¡á¼¹ÀÒ¾à·ÃÅÅÔÊ
ã¹ÃÙ»·Õ 4.11 ãËé¾Ô¨ÒóÒÃÐÂзÕ
k (k th
à·ÃÅÅÔÊ ¨Ð¾ºÇèÒ ÁÕ ¡ÒÃàÊé¹·Ò§¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð 2 àÊé¹·Ò§·Õ ÁÒ¶Ö§ ʶҹР¹Ñ ¹¤×Í
(1, 2)
λk (3, 2)
áÅÐ
(3, 2)
stage) ¢Í§á¼¹ÀÒ¾
(2)
³ àÇÅÒ
ãËé ·Ó¡Òäӹdz¤èÒ àÁµÃÔ¡ ÊÒ¢Ò·Ñ § 2 àÊé¹·Ò§ ¹Ñ ¹¤×Í
k+1
λk (1, 2)
áÅÐ
µÒÁ¢Ñ ¹µÍ¹·Õ (A 4) ¨Ò¡¹Ñ ¹ ʶҹÐàÃÔ Áµé¹·Õ ÊÍ´¤¤Åéͧ¡ÑºàÊé¹·Ò§¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð·Õ ´Õ
·Õ ØÊØ´·Õ ÁҶ֧ʶҹР(2) ³ àÇÅÒ
k+1
¨Ð¶Ù¡àÅ×Í¡µÒÁ¢Ñ ¹µÍ¹·Õ (A 5) ÊÁÁصÔÇèÒ
(1, 2)
¤×Í àÊé¹·Ò§
80
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð·Õ ´Õ·Õ ØÊØ´·Õ ÁҶ֧ʶҹР(2) ³ àÇÅÒ ¹Ñ ¹ àÁµÃÔ¡àÊé¹·Ò§·Õ ÁҶ֧ʶҹР(2) ³ àÇÅÒ áÅÐàÊé¹·Ò§·Õ Âѧ ÁÕ ªÕÇÔµ ÍÂÙè
k+1
´Ñ§¹Ñ ¹¨Ðä´éÇèÒ
k + 1, Φk+1 (2),
πk+1 (2) = 1
ËÅѧ¨Ò¡
¨Ð¶Ù¡»ÃѺ¤èÒµÒÁ¢Ñ ¹µÍ¹·Õ (A 6)
(survivor path) ·Õ ÁÒ¶Ö§ ʶҹР(2) ³ àÇÅÒ
k + 1, Sk+1 (2),
¨Ð¶Ù¡
»ÃѺ¤èÒµÒÁ¢Ñ ¹µÍ¹·Õ (A 7) ãËé·ÓµÒÁ¢Ñ ¹µÍ¹µèÒ§æ àËÅèÒ¹Õ µÒÁÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ仨¹ÊÔ ¹ÊØ´ÅӴѺ ¢éÍÁÙÅ
{yk }
·Õ ä´éÃѺÁÒ áÅÐ¢Ñ ¹µÍ¹ÊØ´·éÒ¡ç¤×Í ¡ÒõѴÊԹ㨨ж١¡ÃзÓâ´Â¡ÒÃàÅ×Í¡àÊé¹·Ò§·Õ ÂѧÁÕ
ªÕÇÔµÍÂÙè·Õ ÁÕ¤èÒàÁµÃÔ¡àÊé¹·Ò§ ³ àÇÅÒ
4.3.4
L + ν , ΦL+ν ,
¹éÍÂ·Õ ÊØ´
¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
à¹× ͧ¨Ò¡ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨Ð·Ó¡ÒûÃÐÁÇżÅÅӴѺ ¢éÍÁÙÅ ·Ñ §ËÁ´¡è͹·Õ ¨ÐµÑ´ÊÔ¹ã¨ÇèÒ ÅӴѺ ¢éÍÁÙÅ·Õ ä´éÃѺ¤ÇèÐà» ¹ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµã´ÁÒ¡·Õ ÊØ´ ´Ñ§¹Ñ ¹ ¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨Ö§¢Ö ¹ÍÂÙè¡ÑºËÅÒ» ¨¨Ñ ´Ñ§¹Õ
1) ¨Ó¹Ç¹¤èÒ·Õ à» ¹ä»ä´é·Ñ §ËÁ´¢Í§¢éÍÁÙÅÍÔ¹¾Øµ
2) ¤ÇÒÁÂÒǢͧÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
3) ˹èǤÇÒÁ¨Ó¢Í§ªèͧ·ÒÃìà¡çµ
|A|
L
ν
ËÃ×ÍÍÒ¨¨ÐÊÃØ»ä´éÇèÒǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔẺ·Õ ãªé§Ò¹¡Ñ¹·Ñ Ç仨ÐÁÕ¨Ó¹Ç¹Ê¶Ò¹Ð·Ñ §ËÁ´ã¹á¼¹ÀÒ¾ à·ÃÅÅÔÊà·èҡѺ ¹éÍÂà·èҡѺ
|A|ν
áÅеéͧ¡ÒÃãªé¨Ó¹Ç¹Ë¹èǤÇÒÁ¨Ó㹡ÒÃà¡çº¢éÍÁÙŵèÒ§æ àªè¹ ¤èÒ
(L + 1)|A|ν
{πk }
ÍÂèÒ§
˹èÇÂ
¨ÐàËç¹ä´éÇèÒ Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔẺ·Õ ãªé§Ò¹¡Ñ¹·Ñ Çä»äÁèÊÒÁÒö¹ÓÁÒãªé§Ò¹¨ÃÔ§ä´éã¹·Ò§»¯ÔºÑµÔ à¹× ͧ¨Ò¡µéͧ¡ÒÃ˹èǤÇÒÁ¨Óà» ¹¨Ó¹Ç¹ÁÒ¡ à¾ÃÒÐ©Ð¹Ñ ¹ÇÔ¸Õ¡ÒÃÅ´¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒ ÇÕà·ÍÃìºÔ ¤×Í¡ÒÃãªé¾ÒÃÒÁÔàµÍÃì·Õ àÃÕ¡ÇèÒ ¤ÇÒÁÅÖ¡¡ÒöʹÃËÑÊ ÇÕà·ÍÃìºÔ àÁ× Í
T
dT
(decoding depth) ã¹ÍÑÅ¡ÍÃÔ·ÖÁ
¤×Í ¤ÒºàÇÅҢͧºÔµ (bit period) ¡ÅèÒǤ×Í Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¨Ð·Ó¡ÒõѴÊÔ¹ã¨
¢éÍÁÙÅ ·ÕÅкԵ ËÅѧ¨Ò¡·Õ àÇÅÒ¼èÒ¹ä»
dT
˹èÇ ´Ñ§¹Ñ ¹ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ·Õ ÁÕ ¡ÒÃãªé ¤ÇÒÁÅÖ¡ ¡ÒÃ
¶Í´ÃËÑʨеéͧ¡Òèӹǹ˹èǤÇÒÁ¨Ó㹡ÒÃà¡çº¢éÍÁÙŵèÒ§æ ÍÂèÒ§¹éÍÂà·èҡѺ ·Ñ Çä»áÅéÇÁÑ¡¨Ðãªé
d ≥ 5(ν + 1)
[27]
(d+1)|A|ν
«Ö §â´Â
4.4.
µÑÇÍÂèÒ§¡ÒÃãªé§Ò¹Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
81
0
(a) 0
0
1
ak = 1
0.5
(b) 1
ak = 0
1
1.5
ÃÙ»·Õ 4.13: á¼¹ÀÒ¾à·ÃÅÅÔʢͧªèͧÊÑÒ³
4.4
H(D) = 1 + 0.5D
µÑÇÍÂèÒ§¡ÒÃãªé§Ò¹Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
ã¹Êèǹ¹Õ ¨ÐáÊ´§µÑÇÍÂèÒ§¢Ñ ¹µÍ¹¡ÒöʹÃËÑÊ¢éÍÁÙÅ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔÍÂèÒ§ÅÐàÍÕ´ ´Ñ§µèÍ仹Õ
µÑÇÍÂèÒ§·Õ 4.2
{0,
¨Ò¡áºº¨ÓÅͧªèͧÊÑÒ³ã¹ÃÙ» ·Õ 4.9 ¶éÒ ¡Ó˹´ãËé ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ
0, 1}, ªèͧÊÑÒ³
hk = δk + 0.5δk−1
delta function), ÊÑҳú¡Ç¹ ¢éÍÁÙÅ
ÇÔ¸Õ·Ó
{yk }
{nk }
=
àÁ× Í
{0.2,
δk
{ak }
=
¤×Í ¿ §¡ìªÑ¹â¤Ã๤à¡ÍÃìà´ÅµÒ (Kronecker
0.5, 0,
−0.35}
¨§áÊ´§¢Ñ ¹µÍ¹¡ÒöʹÃËÑÊ
´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
¨Ò¡·Õ ⨷Âì¡Ó˹´ ¢éÍÁÙÅàÍÒµì¾ØµªèͧÊÑÒ³
rk
ËÒä´é¨Ò¡
rk = ak ∗ hk = {r0 , r1 , r2 , r3 } = {0, 0, 1, 0.5} àÁ× Í
∗
¤×Í µÑÇ´Óà¹Ô¹¡Òä͹âÇÅ٪ѹ (convolution operator) áÅÐ
yk = rk + nk = {0.2, 0.5, 1, 0.15} ¨Ò¡¹Ñ ¹ ãËéÊÃéҧἹÀÒ¾à·ÃÅÅÔʨҡªèͧÊÑÒ³
hk = δk + 0.5δk−1
¹Ñ ¹¤×Í
H(D) = 1 + 0.5D
«Ö § ¨Ð ä´é µÒÁ ÃÙ» ·Õ 4.13 â´Â ã¹ ·Õ ¹Õ á¼¹ÀÒ¾ à· ÃÅÅÔ ÊÁÕ ·Ñ §ËÁ´ 2 ʶҹР¤×Í Ê¶Ò¹Ð (a) áÅРʶҹР(b) ¢Ñ ¹µÍ¹¡ÒöʹÃËÑÊ¢éÍÁÙÅÊÒÁÒöáºè§ÍÍ¡à» ¹ªèǧàÇÅÒµèÒ§æ ´Ñ§áÊ´§ã¹ÃÙ»·Õ 4.14 «Ö § ÁÕÃÒÂÅÐàÍÕ´´Ñ§¹Õ
82
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
0
0.04
Φ1 ( a ) = 0.04
0.64 0.09 0
1.69
Φ1 ( b ) = 0.64 0.04
0.25
Φ ( a ) = 0.29 2
0.25 0 0.64
1
Φ ( b ) = 0.29 2
0.29
1
0 0.25 0.29
0.25
Φ3 ( a ) = 0.54
Φ3 ( b ) = 0.29 0.54
0.02
Φ 4 ( a ) = 0.41
0.72 0.12 0.29
y 0 = 0.2
y1 = 0.5
y2 = 1
1.82
Φ 4 ( b ) = 1.26
y3 = 0.15
ÃÙ»·Õ 4.14: á¼¹ÀҾ͸ԺÒ¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔã¹áµèÅЪèǧàÇÅÒ
4.4.
µÑÇÍÂèÒ§¡ÒÃãªé§Ò¹Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
ªèǧàÇÅÒ·Õ 0
àÁ× Í àÃÔ Áµé¹ ÃѺ ¢éÍÁÙÅ
y0 = 0.2
83
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨Ð·Ó¡ÒáÓ˹´¤èÒ àÃÔ Áµé¹ ¢Í§
àÁµÃÔ¡àÊé¹·Ò§ãËéà·èҡѺ¤èÒÈÙ¹Âì µÒÁ¢Ñ ¹µÍ¹·Õ (A 1) ã¹ÃÙ»·Õ 4.12 ¹Ñ ¹¤×Í
Φ0 (a) = 0
áÅÐ
Φ0 (b) = 0
¨Ò¡¹Ñ ¹ ¡ç¨Ð·Ó¡ÒäӹdzàÁµÃÔ¡ÊÒ¢Ò·Ø¡àÊé¹ÊÒ¢ÒµÒÁ¢Ñ ¹µÍ¹·Õ (A 4) ã¹ÃÙ»·Õ 4.12 ´Ñ§¹Õ
λ0 (a, a) = |0.2 − 0|2 = 0.04 λ0 (a, b) = |0.2 − 1|2 = 0.64 λ0 (b, a) = |0.2 − 0.5|2 = 0.09 λ0 (b, b) = |0.2 − 1.5|2 = 1.69 µÒÁ·Õ à¢Õ¹áÊ´§äÇéã¹áµèÅÐàÊé¹ÊÒ¢Òã¹ÃÙ»·Õ 4.14 ¨Ò¡¹Ñ ¹ ·Õ áµèÅШشµèÍ¡ç¨Ð·Ó¡ÒÃàÅ×Í¡àÊé¹·Ò§¡Òà à»ÅÕ Â¹Ê¶Ò¹Ð·Õ ´Õ·Õ ÊØ´ áÅéÇ¡ç·Ó¡ÒûÃѺ¤èÒàÁµÃÔ¡àÊé¹·Ò§ µÒÁ¢Ñ ¹µÍ¹·Õ (A 5) áÅÐ (A 6) ã¹ÃÙ»·Õ 4.12 ¹Ñ ¹¤×Í
Φ1 (a) = min{0 + 0.04, 0 + 0.09} = 0.04 Φ1 (b) = min{0 + 0.64, 0 + 1.69} = 0.64 µÒÁ·Õ à¢Õ¹áÊ´§äÇé㹨شµèÍã¹ÃÙ»·Õ 4.14 â´Â·Õ àÊé¹ÅÙ¡ÈÃÊÕ´Ó ¤×Í àÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè (survivor path) ÊèǹàÊé¹ÅÙ¡ÈÃÊÕà·Ò ¤×Í àÊé¹·Ò§·Õ ¶Ù¡µÑ´·Ô §
ªèǧàÇÅÒ·Õ 1
àÁ× Í ÃѺ ¢éÍÁÙÅ
y1 = 0.5
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨Ð·Ó¡ÒäӹdzàÁµÃÔ¡ ÊÒ¢Ò·Ñ §ËÁ´
´Ñ§¹Õ
λ1 (a, a) = |0.5 − 0|2 = 0.25 λ1 (a, b) = |0.5 − 1|2 = 0.25 λ1 (b, a) = |0.5 − 0.5|2 = 0 λ1 (b, b) = |0.5 − 1.5|2 = 1
84
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
µÒÁ·Õ à¢Õ¹áÊ´§äÇéã¹áµèÅÐàÊé¹ÊÒ¢Òã¹ÃÙ»·Õ 4.14 ¨Ò¡¹Ñ ¹ ·Õ áµèÅШشµèÍ¡ç¨Ð·Ó¡ÒÃàÅ×Í¡àÊé¹·Ò§¡Òà à»ÅÕ Â¹Ê¶Ò¹Ð·Õ ´Õ·Õ ÊØ´ áÅéÇ¡ç·Ó¡ÒûÃѺ¤èÒàÁµÃÔ¡àÊé¹·Ò§ ¹Ñ ¹¤×Í
Φ2 (a) = min{0.04 + 0.25, 0.64 + 0} = 0.29 Φ2 (b) = min{0.04 + 0.25, 0.64 + 1} = 0.29 µÒÁ·Õ à¢Õ¹áÊ´§äÇé㹨شµèÍã¹ÃÙ»·Õ 4.14
ªèǧàÇÅÒ·Õ 2
àÁ× ÍÃѺ¢éÍÁÙÅ
y2 = 1
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¨Ð·Ó¡ÒäӹdzàÁµÃÔ¡ÊÒ¢Ò·Ñ §ËÁ´ ´Ñ§¹Õ
λ2 (a, a) = |1 − 0|2 = 1 λ2 (a, b) = |1 − 1|2 = 0 λ2 (b, a) = |1 − 0.5|2 = 0.25 λ2 (b, b) = |1 − 1.5|2 = 0.25 µÒÁ·Õ à¢Õ¹áÊ´§äÇéã¹áµèÅÐàÊé¹ÊÒ¢Òã¹ÃÙ»·Õ 4.14 ¨Ò¡¹Ñ ¹ ·Õ áµèÅШشµèÍ¡ç¨Ð·Ó¡ÒÃàÅ×Í¡àÊé¹·Ò§¡Òà à»ÅÕ Â¹Ê¶Ò¹Ð·Õ ´Õ·Õ ÊØ´ áÅéÇ¡ç·Ó¡ÒûÃѺ¤èÒàÁµÃÔ¡àÊé¹·Ò§ ¹Ñ ¹¤×Í
Φ3 (a) = min{0.29 + 1, 0.29 + 0.25} = 0.54 Φ3 (b) = min{0.29 + 0, 0.29 + 0.25} = 0.29 µÒÁ·Õ à¢Õ¹áÊ´§äÇé㹨شµèÍã¹ÃÙ»·Õ 4.14
ªèǧàÇÅÒ·Õ 3
àÁ× Í ÃѺ ¢éÍÁÙÅ
y3 = 0.15
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨Ð·Ó¡ÒäӹdzàÁµÃÔ¡ ÊÒ¢Ò·Ñ §ËÁ´
´Ñ§¹Õ
λ3 (a, a) = |0.15 − 0|2 = 0.02 λ3 (a, b) = |0.15 − 1|2 = 0.72 λ3 (b, a) = |0.15 − 0.5|2 = 0.12 λ3 (b, b) = |0.15 − 1.5|2 = 1.82
4.4.
µÑÇÍÂèÒ§¡ÒÃãªé§Ò¹Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
85
µÒÁ·Õ à¢Õ¹áÊ´§äÇéã¹áµèÅÐàÊé¹ÊÒ¢Òã¹ÃÙ»·Õ 4.14 ¨Ò¡¹Ñ ¹ ·Õ áµèÅШشµèÍ¡ç¨Ð·Ó¡ÒÃàÅ×Í¡àÊé¹·Ò§¡Òà à»ÅÕ Â¹Ê¶Ò¹Ð·Õ ´Õ·Õ ÊØ´ áÅéÇ¡ç·Ó¡ÒûÃѺ¤èÒàÁµÃÔ¡àÊé¹·Ò§ ¹Ñ ¹¤×Í
Φ4 (a) = min{0.54 + 0.02, 0.29 + 0.12} = 0.41 Φ4 (b) = min{0.54 + 0.72, 0.29 + 1.82} = 1.26 µÒÁ·Õ à¢Õ¹áÊ´§äÇé㹨شµèÍã¹ÃÙ»·Õ 4.14
ËÅѧ¨Ò¡·ÓµÒÁ¢Ñ ¹µÍ¹¢Í§ÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ¨¹ÊÔ ¹ÊØ´ÅӴѺ¢éÍÁÙÅ·Õ ä´éÃѺ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
â
¡ç ¨Ð·Ó¡ÒöʹÃËÑÊ ¢éÍÁÙÅ ÍÔ¹¾Øµ ¨Ðä´éÇèÒ
Φ4 (a) = 0.41
¨Ò¡àÊé¹·Ò§·Õ Âѧ ÁÕ ªÕÇÔµ ÍÂÙè ·Õ ÁÕ ¤èÒ àÁµÃÔ¡ àÊé¹·Ò§¹éÍÂ·Õ ÊØ´ ã¹·Õ ¹Õ
ÁÕ¤èÒ¹éÍÂ·Õ ÊØ´ à¾ÃÒÐ©Ð¹Ñ ¹ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¨ÐÁͧÂé͹¡ÅÑºä» (trace
back) µÒÁàÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè·Õ ÁÒ¶Ö§¨Ø´µèÍ
Φ4 (a)
¡ç¨Ð¾ºÇèÒ ¢éÍÁÙÅÍÔ¹¾Øµ
{âk }
·Õ ÊÍ´¤Åéͧ¡Ñº
àÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè¹Õ ¤×Í
{âk } = {â0 , â1 , â2 } = {0, 0, 1} «Ö §µÃ§¡Ñº¢éÍÁÙÅÍÔ¹¾Øµ
{ak }
·Õ Êè§ÁÒ¨Ò¡µé¹·Ò§¨ÃÔ§ áÊ´§ÇèÒ¡ÒöʹÃËÑÊ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¹Õ
äÁèÁÕ¢éͼԴ¾ÅÒ´à¡Ô´¢Ö ¹ã¹Ãкº ËÁÒÂà˵Ø
−1)
¢éÍÁÙÅ ÍÔ¹¾Øµ µÑÇ ÊØ´·éÒÂ·Õ ÊÒÁÒö¶Í´ÃËÑÊ ä´é ¨Ò¡Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ (ã¹·Õ ¹Õ ¤×Í
a3 =
äÁè¨Óà» ¹µéͧ¶Í´ÃËÑÊ à¾ÃÒж×ÍÇèÒ à» ¹¢éÍÁÙźԵÊèǹà¡Ô¹·Õ ä´é¨Ò¡¡Ò÷Ӥ͹âÇÅ٪ѹ¢Í§¢éÍÁÙÅ
ÍÔ¹¾Øµ ¡Ñº ªèͧÊÑÒ³ áµè ¢éÍÁÙÅ
y3
à» ¹ ¢éÍÁÙÅ ·Õ ¨Óà» ¹ ·Õ ¨Ðµéͧ¹ÓÁÒãªé 㹡ÒöʹÃËÑÊ ¢éÍÁÙÅ ¢Í§
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
µÑÇÍÂèÒ§·Õ 4.3
{−1, −1,
1,
¨Ò¡áºº¨ÓÅͧªèͧÊÑÒ³ã¹ÃÙ» ·Õ 4.9 ¶éÒ ¡Ó˹´ãËé ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ
−1},
ÊÑҳú¡Ç¹
{nk }
ªèͧÊÑÒ³ =
hk = δk − δk−1
{0.5, −0.4,
0.1, 0.7,
¡) ÇÒ´á¼¹ÀÒ¾ºÅçÍ¡¢Í§ªèͧÊÑÒ³
−0.3}
àÁ× Í
δk
{ak }
=
¤×Í ¿ §¡ìªÑ¹ â¤Ã๤à¡ÍÃìà´ÅµÒ,
¨§
hk
¢) à¢Õ¹ÊÁ¡ÒÃáÊ´§¤ÇÒÁÊÑÁ¾Ñ¹¸ì ÃÐËÇèÒ§¢éÍÁÙÅ ÍÔ¹¾Øµ áÅТéÍÁÙÅ àÍÒµì¾Øµ ¢Í§ªèͧÊÑÒ³ áÅÐ ¤Ó¹Ç³ËÒ¤èÒ
{yk }
86
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
rk ak
0 2
-1
D
(a) Block diagram
-2
1
1/2
--1 1-
-1/0
-1
111
1
0
ak = 1 ak = -1
1/0
(c) Trellis diagram -1/-2
(b) FSM ÃÙ»·Õ 4.15: (a) á¼¹ÀÒ¾ºÅçÍ¡, (b) à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´, áÅÐ (c) á¼¹ÀÒ¾à·ÃÅÅÔÊ ¢Í§ªèͧÊÑÒ³
H(D) = 1 − D
¤) ÊÃéÒ§à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ áÅÐÇÒ´á¼¹ÀÒ¾à·ÃÅÅÔÊ ¢Í§ªèͧÊÑÒ³
§) ¶Í´ÃËÑÊ¢éÍÁÙÅ
{yk }
hk
´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
ÇÔ¸Õ·Ó ¡)
¨Ò¡ªèͧÊÑÒ³·Õ ¡Ó˹´ÁÒãËé
hk = δk − δk−1
¹Ñ ¹¤×Í
H(D) = 1 − D
ÊÒÁÒöáÊ´§à» ¹
á¼¹ÀÒ¾ºÅçÍ¡¢Í§ªèͧÊÑÒ³ä´éµÒÁÃÙ»·Õ 4.15(a) ¢)
à¹× ͧ¨Ò¡
àÍÒµì¾Øµ
rk
H(D) = 1 − D
´Ñ§¹Ñ ¹ ÊÁ¡ÒÃáÊ´§¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§¢éÍÁÙÅÍÔ¹¾Øµ
ak
áÅТéÍÁÙÅ
¢Í§ªèͧÊÑÒ³¹Õ ¤×Í
rk = ak − ak−1 «Ö §¨Ðä´éÇèÒ
rk
=
{−1,
0, 2,
−2, 1}
´Ñ§¹Ñ ¹
yk = rk + nk
=
{−0.5, −0.4,
2.1,
−1.3,
0.7}
4.4.
µÑÇÍÂèÒ§¡ÒÃãªé§Ò¹Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
0
-1
0.25
0.25
0.41
0.16
87
4.41
0.91
4.82 0.49
1
0.25
0.25
0.16
0.41
0.42
1.69
1.40
1.69
0.01
0
0.49
2.11
0.49
2.60
ÃÙ»·Õ 4.16: á¼¹ÀÒ¾ÊÃØ»¢Ñ ¹µÍ¹¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
¤)
¨Ò¡ÊÁ¡ÒÃáÊ´§¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§¢éÍÁÙÅÍÔ¹¾ØµáÅТéÍÁÙÅàÍÒµì¾Øµ·Õ ä´éã¹¢éÍ ¢) à¤Ã× Í§Ê¶Ò¹Ð
¨Ó¡Ñ´áÅÐá¼¹ÀÒ¾à·ÃÅÅÔÊ ÊÒÁÒöáÊ´§ä´éµÒÁÃÙ»·Õ 4.15(b) áÅÐ 4.15(c) µÒÁÅӴѺ §)
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔãªéá¼¹ÀÒ¾à·ÃÅÅÔÊ㹡ÒöʹÃËÑÊ¢éÍÁÙÅ
{yk } â´Â·Õ ¢Ñ ¹µÍ¹¡ÒöʹÃËÑÊ
¢éÍÁÙÅ ÊÒÁÒöÊÃØ» ä´é µÒÁÃÙ» ·Õ 4.16 àÁ× Í µÑÇàÅ¢·Õ áÊ´§ÍÂÙè º¹¨Ø´µèÍ áµèÅШش ¤×Í ¤èÒ àÁµÃÔ¡ àÊé¹·Ò§ ·Õ ÁÒ¶Ö§ ³ ¨Ø´µèÍ ¹Ñ ¹ áÅеÑÇàÅ¢·Õ áÊ´§ÍÂÙè º¹àÊé¹ ÊÒ¢ÒáµèÅÐàÊé¹ ¤×Í ¤èÒ àÁµÃÔ¡ ÊҢҢͧáµèÅÐàÊé¹ ÊÒ¢Ò·Õ ´Õ ·Õ ÊØ´ ·Õ ÁÒ¶Ö§ ·Õ ¨Ø´µèÍ ¹Ñ ¹æ ¨Ò¡ÃÙ» ·Õ 4.16 ¤èÒ àÁµÃÔ¡ àÊé¹·Ò§·Õ ¹éÍÂ·Õ ÊØ´ ¤×Í ¤èÒ 1.40 ´Ñ§¹Ñ ¹ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¨Ð¶Í´ÃËÑÊ¢éÍÁÙÅâ´Â¡ÒÃÁͧÂé͹¡ÅѺ仵ÒÁàÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè·Õ ÁÒ¶Ö§¨Ø´µèÍ ·Õ ÁÕ¤èÒàÁµÃÔ¡àÊé¹·Ò§à·èҡѺ 1.40 «Ö §¨Ð¾ºÇèÒ ¤èÒ»ÃÐÁÒ³¢Í§ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
{âk }
·Õ ÊÍ´¤Åéͧ
¡ÑºàÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè¹Õ ¤×Í
{âk } = {â0 , â1 , â2 , â3 } = {−1, −1, 1, −1} «Ö §ÁÕ¤èҵç¡ÑºÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
{ak }
·Õ Êè§ÁҨҡǧ¨ÃÀÒ¤Êè§ à¾ÃÒÐ©Ð¹Ñ ¹ ¡ÒöʹÃËÑÊ´éÇÂǧ¨Ã
µÃǨËÒÇÕà·ÍÃìºÔã¹µÑÇÍÂèÒ§¢éÍ¹Õ ¨Ö§äÁèÁÕ¢éͼԴ¾ÅÒ´à¡Ô´¢Ö ¹
4.4.1
ÊÃػǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨Ð·Ó§Ò¹ä´é ÍÂèÒ§ÁÕ »ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡·Õ ÊØ´ ¡ç µèÍàÁ× Í Í§¤ì»ÃСͺ¢Í§ÊÑÒ³ ú¡Ç¹·Õ ὧÍÂÙèã¹¢éÍÁÙÅ·Õ ¨Ð·Ó¡ÒöʹÃËÑÊ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔÁÕÅѡɳÐà» ¹ÊÑҳú¡Ç¹ à¡ÒÊì ÊÕ ¢ÒÇẺºÇ¡ (AWGN) [15] ·Ñ §¹Õ à» ¹ à¾ÃÒÐÇèÒ ÍÑÅ¡ÍÃÔ·ÖÁ ÇÕà·ÍÃìºÔ ä´é ÁÒ¨Ò¡ÊÁÁص԰ҹ·Õ ÇèÒ
88
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
µÒÃÒ§·Õ 4.2: µÑÇÍÂèÒ§áÊ´§¨Ó¹Ç¹Ê¶Ò¹Ð·Õ µéͧãªéã¹á¼¹ÀÒ¾à·ÃÅÅÔʢͧ·ÒÃìà¡çµáººµèÒ§æ
H(D)
·ÒÃìà¡çµáºº PR PR4
[1 0
EPR4 EEPR4
−1]
[1 1
−1 −1]
[1 2 0
−2 −1]
˹èǤÇÒÁ¨Ó
ν
¨Ó¹Ç¹Ê¶Ò¹Ð·Ñ §ËÁ´
(1 − D)(1 + D)
2
22 = 4
(1 − D)(1 + D)2
3
23 = 8
(1 − D)(1 + D)3
4
24 = 16
ͧ¤ì»ÃСͺ¢Í§ÊÑҳú¡Ç¹·Õ ´éÒ¹¢Òà¢éҢͧǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔà» ¹ÊÑҳú¡Ç¹áººà¡ÒÊì ÊÕ¢ÒÇẺºÇ¡ ˹éÒ·Õ ËÅÑ¡ ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¤×Í ¨Ð·Ó¡ÒöʹÃËÑÊ ¢éÍÁÙÅ â´ÂãËé ÁÕ ¤èÒ¤ÇÒÁ¹èÒ¨Ðà» ¹ ¢Í§ ¢éÍ ¼Ô´¾ÅÒ´ÅӴѺ (probability of sequence error) ¹éÍÂ·Õ ÊØ´ áÅÐâ´Â·Ñ Ç令ÇÒÁ«Ñº«é͹¢Í§Ç§¨Ã µÃǨËÒÇÕà·ÍÃìºÔ¨Ð¢Ö ¹ÍÂÙè¡Ñº¨Ó¹Ç¹Ê¶Ò¹Ð·Õ ãªéã¹á¼¹ÀÒ¾à·ÃÅÅÔÊ ´Ñ§áÊ´§ã¹µÒÃÒ§·Õ 4.2 ¨Ð¾ºÇèÒ àÁ× Í ·ÒÃìà¡çµ ·Õ ãªé ÁÕ ¨Ó¹Ç¹á·ç» (ËÃ×Í Ë¹èǤÇÒÁ¨Ó) ÁÒ¡¢Ö ¹ ¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¡ç ¨ÐÁÒ¡¢Ö ¹ à¹× ͧ¨Ò¡ ¨Ó¹Ç¹Ê¶Ò¹Ð·Õ ãªé ã¹á¼¹ÀÒ¾à·ÃÅÅÔ ÊÁÕ ÁÒ¡¢Ö ¹ ÍÂèÒ§äáçµÒÁ àÁ× Í ¤ÇÒÁ¨Ø ¢éÍÁÙŢͧÎÒÃì´´ÔÊ¡ìä´Ã¿ìÊÙ§¢Ö ¹ ·ÒÃìà¡çµ·Õ ¨Ð¹ÓÁÒãªé¡ç¤ÇÃ·Õ ¨ÐµéͧÁըӹǹá·ç»ÁÒ¡¢Ö ¹ à¾× Í·ÓãËé¼Å µÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§·ÒÃìà¡çµ ÊÍ´¤Åéͧ¡Ñº ¼ÅµÍºÊ¹Í§àªÔ§ ¤ÇÒÁ¶Õ ¢Í§ªèͧÊÑÒ³ ´Ñ§¹Ñ ¹ã¹ ¡ÒþԨÒóÒÇèÒ ¨Ð¹Ó·ÒÃìà¡çµ ·Õ ÁÕ ¨Ó¹Ç¹á·ç» ÁÒ¡ÁÒãªé §Ò¹ËÃ×Í äÁè ¹Ñ ¹ ¨Ðµéͧ»ÃйջÃйÍÁÃÐËÇèÒ§ »ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº·Õ ¨Ðä´éÃѺ áÅФÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
4.5
ÊÃØ»·éÒº·
à·¤¹Ô¤ PRML à» ¹ ¡ÒÃãªé §Ò¹ÃèÇÁ¡Ñ¹ ÃÐËÇèÒ§ÍÕ¤ÇÍäÅà«ÍÃì Ẻ PR áÅÐǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ «Ö § à» ¹·Õ ¹ÔÂÁãªé§Ò¹¡Ñ¹ÁÒ¡ã¹Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì ÊÒà˵ØË¹Ö §ÍÒ¨¨Ðà» ¹ à¾ÃÒÐÇèÒǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¶×Íä´éÇèÒà» ¹Ç§¨ÃµÃǨËÒ¢éÍÁÙÅ·Õ ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡ â´Â੾ÒÐÍÂèÒ§ÂÔ § àÁ× Í Í§¤ì »ÃСͺ¢Í§ÊÑ Ò³Ãº¡Ç¹·Õ ´éÒ¹¢Òà¢éÒ ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÁÕ ÅѡɳÐà» ¹ ÊÑÒ³ ú¡Ç¹áººà¡ÒÊì ÊÕ ¢ÒÇẺºÇ¡ ËÅÑ¡ ¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨ÐÍÂÙè º¹¾× ¹ °Ò¹¢Í§á¼¹ ÀÒ¾à·ÃÅÅÔÊ«Ö §ÍÒ¨¨ÐÊÃéÒ§ä´é¨Ò¡à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ áÅÐâ´Â·Ñ Çä»áÅéǤÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒ
4.6.
à຺½ ¡ËÑ´·éÒº·
89
ÇÕà·ÍÃìºÔ¨Ð¢Ö ¹ÍÂÙè¡Ñº¨Ó¹Ç¹Ê¶Ò¹Ð·Ñ §ËÁ´·Õ µéͧãªéã¹á¼¹ÀÒ¾à·ÃÅÅÔÊ «Ö §¨Ó¹Ç¹Ê¶Ò¹Ð·Õ ãªéã¹á¼¹ ÀÒ¾à·ÃÅÅÔÊ¨Ð¢Ö ¹ÍÂÙè¡Ñº ¨Ó¹Ç¹¤èÒ·Õ à» ¹ä»ä´é·Ñ §ËÁ´¢Í§¢éÍÁÙÅÍÔ¹¾Øµ áÅÐ˹èǤÇÒÁ¨Ó¢Í§·ÒÃìà¡çµ ´Ñ§¹Ñ ¹ ¡ÒèйӷÒÃìà¡çµ ã´ÁÒãªé §Ò¹ã¹ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ¨Ðµéͧ»ÃйջÃйÍÁÃÐËÇèÒ§»ÃÐÊÔ·¸ÔÀÒ¾ ¢Í§Ãкº·Õ ¨Ðä´éÃѺ áÅФÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
4.6
à຺½ ¡ËÑ´·éÒº·
1. ¨§Í¸ÔºÒ¢éÍᵡµèÒ§ÃÐËÇèÒ§ÍÕ¤ÇÍäÅà«ÍÃìẺ¼ÅµÍºÊ¹Í§àµçÁ (full response) áÅÐẺ¼Å µÍºÊ¹Í§ºÒ§Êèǹ (partial response) 2. ¡Ó˹´ãËé¢éÍÁÙÅÍÔ¹¾Øµ
ak ∈ {−1, 1}
¨§ÇÒ´á¼¹ÀÒ¾ºÅçÍ¡ (block diagram), à¤Ã× Í§Ê¶Ò¹Ð
¨Ó¡Ñ´ (FSM), áÅÐá¼¹ÀÒ¾à·ÃÅÅÔÊ (Trellis diagram) ¢Í§ªèͧÊÑÒ³ 2.1)
H(D) = 1 − 0.5D
2.2)
H(D) = 1 + 2D + D2
2.3)
H(D) = 1 − D3
2.4)
H(D) = 1 + 3D + 3D2 + D3
3. ¨Ò¡áºº¨ÓÅͧªèͧÊÑÒ³ã¹ÃÙ» ·Õ 4.9 ¶éÒ ¡Ó˹´ãËé ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ
−1,
1} ¶Ù¡ Êè§ ¼èÒ¹ªèͧÊÑÒ³
−0.4,
0.1, 0.7,
−0.3,
H(D)
H(D)
{ak }
«Ö § ¶Ù¡ ú¡Ç¹´éÇÂÊÑҳú¡Ç¹
0.4} ¨§¶Í´ÃËÑÊ¢éÍÁÙÅ
{yk }
µèÍ仹Õ
{1, −1,
=
{nk }
=
{0.5,
â´Âãªéǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÊÓËÃѺªèͧ
ÊÑÒ³µèÍä»¹Õ 3.1)
H(D) = 1 + 2D + D2
3.2)
H(D) = 1 − D2
4. ¨Ò¡áºº¨ÓÅͧªèͧÊÑÒ³ã¹ÃÙ» ·Õ 4.9 ¶éÒ ¡Ó˹´ãËé ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ 0} ¶Ù¡Ê觼èÒ¹ªèͧÊÑÒ³ 0.2, 0.6,
−0.4, −0.3,
ÊÑÒ³µèÍ仹Õ
H(D)
«Ö §¶Ù¡Ãº¡Ç¹´éÇÂÊÑҳú¡Ç¹
0.5} ¨§¶Í´ÃËÑÊ¢éÍÁÙÅ
{yk }
{ak }
{nk }
=
=
{1,
0, 1,
{0.3, −0.5,
â´Âãªéǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÊÓËÃѺªèͧ
90
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
4.1)
H(D) = 1 + 3D + 3D2 + D3
4.2)
H(D) = 1 + D − D2 − D3
º··Õ 5
¡ÒÃÇÔà¤ÃÒÐËìà˵ءÒóì¢éͼԴ¾ÅÒ´
㹺·¹Õ ¨Ð͸ԺÒ¶֧ ËÅÑ¡¡ÒÃÇÔà¤ÃÒÐËì à˵ءÒÃ³ì ¢éͼԴ¾ÅÒ´ (error event) «Ö § à» ¹ ¡Òäé¹ËÒÃٻẺ ¢Í§¢éͼԴ¾ÅÒ´·Õ à¡Ô´ ¢Ö ¹ ºèÍ ã¹ÃÐËÇèÒ§¡Ãкǹ¡ÒöʹÃËÑÊ ¢éÍÁÙÅ ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ áÅÐ àÁ× Í ÊÒÁÒöÇèÒ ¢éͼԴ¾ÅÒ´ÃÙ»áººã´·Õ à¡Ô´ ¢Ö ¹ ºèÍ ¹Ñ¡Í͡ẺÃкº¡ç ÊÒÁÒö·Õ ¨Ð·Ó¡ÒÃÍ͡Ẻ ÃËÑÊ RLL (run length limited)
[9] ËÃ×Í Ç§¨Ãà¢éÒÃËÑÊ¡è͹
(precoder) [47] à¾× Íãªéà¢éÒÃËÑÊ¢éÍÁÙÅ
¢èÒÇÊÒáè͹·Õ ¨Ð·Ó¡ÒÃà¢Õ¹ŧä»ã¹Ê× ÍºÑ¹·Ö¡ à¾× ÍËÅÕ¡àÅÕ Â§¡ÒÃà¡Ô´¢éͼԴ¾ÅÒ´àËÅèÒ¹Ñ ¹ÃÐËÇèÒ§¡Òà ¶Í´ÃËÑÊ ¢éÍÁÙÅ ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ «Ö § ¨ÐÊè§ ¼Å·ÓãËé »ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкºã¹ÃÙ» ¢Í§ÍѵÃÒ¢éÍ ¼Ô´¾ÅÒ´ºÔµ (BER) ¹éÍÂŧÁÒ¡ ¹Í¡¨Ò¡¹Õ ¶éÒ·ÃÒºÇèÒ¢éͼԴ¾ÅÒ´ÃÙ»áººã´·Õ à¡Ô´¢Ö ¹ºèÍÂã¹Ãкº ¡çÊÒÁÒö·Õ ¨Ð¹Ó¢éÍÁÙÅ¹Õ ÁÒãªé㹡ÒÃÇÔà¤ÃÒÐËìáÅÐà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§·ÒÃìà¡çµáººµèÒ§æ ä´é â´Âãªé ¾ÒÃÒÁÔàµÍÃì ·Õ àÃÕ¡ÇèÒ SNR »ÃÐÊÔ·¸Ô¼Å (e ective SNR) [48] «Ö § ÁÕ ¼ÅÅѾ¸ì à·Õºà·èÒ ¡Ñº ¡Òà ãªé¾ÒÃÒÁÔàµÍÃì BER 㹡ÒÃà»ÃÕºà·Õº áµèãªéàÇÅÒ㹡ÒäӹdzËÒ¤èÒ SNR »ÃÐÊÔ·¸Ô¼Å¹Õ ¹éÍ¡ÇèÒ ÁÒ¡ ´Ñ§ÃÒÂÅÐàÍÕ´·Õ ¨Ð͸ԺÒµèÍä»ã¹º·¹Õ
5.1
º·¹Ó
µÒÁ·Õ ͸ԺÒÂã¹ËÑÇ¢éÍ·Õ 4.3 ¾ºÇèÒǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¨Ð·Ó§Ò¹ÍÂÙ躹¾× ¹°Ò¹¢Í§á¼¹ÀÒ¾à·ÃÅÅÔÊ (trellis diagram) «Ö §ã¹¡ÒöʹÃËÑÊ¢éÍÁÙźҧ¤ÃÑ § ÁÕ¤ÇÒÁà» ¹ä»ä´é·Õ ¨Ðà¡Ô´à˵ءÒóìµÒÁÃÙ»·Õ 5.1 91
92
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
B
A
C
E
D F #1
-1
1
-1
1
-1
-1
1
1
#2
-1
1
1
-1
1
-1
1
1
Error 0 sequence
0
-2
2
-2
0
0
0
ÃÙ»·Õ 5.1: µÑÇÍÂèÒ§¡Ò÷ӧҹÀÒÂã¹à·ÃÅÅÔʢͧǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
¹Ñ ¹¤×Í ã¹¢³Ð·Õ ·Ó¡ÒöʹÃËÑÊ ¢éÍÁÙÅ ¼èҹἹÀÒ¾à·ÃÅÅÔÊ ¨¹ÊÔ ¹ÊØ´ ÅӴѺ ¢éÍÁÙÅ ·Õ ä´é ÃѺ »ÃÒ¡®ÇèÒ àÊé¹·Ò§·Õ Âѧ ÁÕ ªÕÇÔµ ÍÂÙè (survivor path) ·Õ ´Õ ·Õ ÊØ´ à» ¹ 仵ÒÁÃÙ» ·Õ 5.1 ¡ÅèÒǤ×Í àÁ× Í ¤Ó¹Ç³ËÒ¤èÒ àÁµÃÔ¡àÊé¹·Ò§ (path metric) ·Õ ¨Ø´ E à¾× Í·Ó¡ÒÃàÅ×Í¡àÊé¹·Ò§·Õ ´Õ·Õ ÊØ´·Õ ÁÒ¶Ö§¨Ø´ E ¾ºÇèÒÁÕàÊé¹·Ò§ 2 àÊé¹·Ò§ ¤×Í àÊé¹·Ò§ #1 (àÊé¹·Ò§ ABCEF) áÅÐàÊé¹·Ò§ #2 (àÊé¹·Ò§ ABDEF) ·Õ ÁÕ¤èÒ àÁµÃÔ¡ àÊé¹·Ò§ ³ ¨Ø´ E à·èҡѹ (àÊé¹·Ò§·Ñ §ÊͧÁÕ¤ÇÒÁà» ¹ä»ä´éà·èҡѹ·Õ ¨Ð¶Ù¡àÅ×Í¡â´ÂÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ) à¾ÃÒÐ©Ð¹Ñ ¹ ã¹¡Ã³Õ ¹Õ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÊÒÁÒö·Õ ¨ÐµÑ´ÊÔ¹ã¨àÅ×Í¡àÊ鹷ҧ㴡çä´é à¾× Í ãªé à» ¹ ÊèÇ¹Ë¹Ö § ¢Í§àÊé¹·Ò§·Õ Âѧ ÁÕ ªÕÇÔµ ÍÂÙè ·Õ ´Õ ·Õ ÊØ´ ÊÓËÃѺ ¡ÒöʹÃËÑÊ ¢éÍÁÙÅ µÑÇÍÂèÒ§àªè¹ ¶éÒ Ç§¨ÃµÃǨËÒ ÇÕà·ÍÃìºÔàÅ×Í¡àÊé¹·Ò§ #1 ¡ç¨Ðä´é¶Í´ÃËÑÊ¢éÍÁÙÅä´éà» ¹ àÊé¹·Ò§ #2 ¡ç¨Ðä´é¼ÅÅѾ¸ìà» ¹ ¨ÃÔ§ ¤×Í
{−1,
¢éÍÁÙÅà» ¹
1,
{−1,
−1, 1, 1,
1,
{−1,
−1, −1,
−1,
1,
−1,
1, 1,
−1,
1,
{−1,
−1,
1,
−1,
1,
−1, −1,
1, 1} áµè¶éÒàÅ×Í¡
1, 1} ´Ñ§¹Ñ ¹¶éÒÊÁÁصÔÇèÒ ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
1, 1} ¡çËÁÒ¤ÇÒÁÇèÒ âÍ¡ÒÊ·Õ Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¨Ð¶Í´ÃËÑÊ 1, 1} ¡çÁÕ¤ÇÒÁà» ¹ä»ä´éÊÙ§ (ËÃ×Íã¹·Ò§¡ÅѺ¡Ñ¹) à¾ÃÒÐ©Ð¹Ñ ¹
¶éÒ¹ÓàÍÒÅӴѺ¢éÍÁÙÅ·Ñ §Êͧ (ÅӴѺ·Õ #1 áÅÐ #2) ÁÒź¡Ñ¹ àªè¹ ÅӴѺ¢éÍÁÙÅ #1 ź´éÇÂÅӴѺ¢éÍÁÙÅ #2 ¡ç ¨Ðä´é ¼ÅÅѾ¸ì à» ¹
{0,
0,
−2,
2,
−2,
0, 0, 0} ËÃ×Í ¶éÒ ¹ÓàÍÒÅӴѺ ¢éÍÁÙÅ #2 ź´éÇÂÅӴѺ
5.2.
¤ÇÒÁËÁÒ¢ͧà˵ءÒóì¢éͼԴ¾ÅÒ´
93
W(D) A(D)
H(D)
X(D)
Y(D)
Viterbi detector
Aˆ ( D )
ÃÙ»·Õ 5.2: Ẻ¨ÓÅͧªèͧÊÑÒ³ GPR ẺÊÁÁÙÅ
¢éÍÁÙÅ #1 ¡ç¨Ðä´é¼ÅÅѾ¸ìà» ¹
{2, −2, 2}
{0,
0, 2,
−2,
2, 0, 0, 0} ´Ñ§¹Ñ ¹ ¼ÅÅѾ¸ì·Õ ä´é
{−2,
2,
−2}
ËÃ×Í
¨ÐàÃÕ¡¡Ñ¹ ÇèÒ ÅӴѺ ¢éͼԴ¾ÅÒ´ (error sequence) [15, 49, 50] «Ö § àÍÒäÇé ãªé 㹡ÒÃ
¤Ó¹Ç³ËÒà˵ءÒóì¢éͼԴ¾ÅÒ´¢Í§Ãкº ¡ÒÃÇÔà¤ÃÒÐËì à˵ءÒÃ³ì ¢éͼԴ¾ÅÒ´¢Í§Ãкº¨ÐªèÇ·ÓãËé ·ÃÒºÇèÒ ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ ÃÙ»áººã´·Õ ¨ÐÊè§ ¼Å¡Ãзº·ÓãËé ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ à¡Ô´ ¤ÇÒÁÊѺʹã¹ÃÐËÇèÒ§·Õ ·Ó¡ÒöʹÃËÑÊ ¢éÍÁÙÅ (à¡Ô´ »ÃÒ¡®¡ÒóìµÒÁÃÙ»·Õ 5.1) ´Ñ§¹Ñ ¹ ¶éÒÊÒÁÒö·Õ ¨ÐËÅÕ¡àÅÕ Â§¡ÒÃÊè§ÅӴѺ¢éÍÁÙÅÍÔ¹¾ØµÃٻẺàËÅèÒ¹Ñ ¹ à¢éÒä»ã¹Ãкºä´é ¡ç¨ÐªèÇ·ÓãËéǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔà¡Ô´¢éͼԴ¾ÅҴ㹡ÒöʹÃËÑÊ¢éÍÁÙŹéÍÂŧ «Ö § ¨ÐÊ觼ŷÓãËé»ÃÐÊÔ·¸ÔÀÒ¾ÃÇÁ¢Í§Ãкº´ÕÂÔ §¢Ö ¹ [49, 50]
5.2
¤ÇÒÁËÁÒ¢ͧà˵ءÒóì¢éͼԴ¾ÅÒ´
Ẻ¨ÓÅͧ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµã¹ÃÙ»·Õ 3.2 ÊÒÁÒö·Õ ¨Ð¨Ñ´ãËéÍÂÙèã¹ÃÙ»¢Í§áºº¨ÓÅͧªèͧÊÑÒ³ GPR (generalized partial response) ẺÊÁÁÙÅ ´Ñ§áÊ´§ã¹ÃÙ»·Õ 5.2 àÁ× Í ÅӴѺ¢éͼԴ¾ÅÒ´ ã¹ÃÙ» ·Õ 3.2 ÊÒÁÒö·Õ ¨Ð¶Ù¡ ¾Ô¨ÒóÒÇèÒ à» ¹ ÊÑҳú¡Ç¹
W (D)
wk
·Õ ´éÒ¹¢Òà¢éÒ ¢Í§Ç§¨ÃµÃǨËÒ
ÇÕà·ÍÃìºÔ µÒÁÃÙ»·Õ 5.2 ã¹·Ò§»¯ÔºÑµÔ (â´Â੾ÒÐÍÂèÒ§ÂÔ § ·Õ ¤ÇÒÁ¨Ø¢éÍÁÙŢͧÎÒÃì´´ÔÊ¡ìä´Ã¿ì ND ÊÙ§) ÊÑҳú¡Ç¹
wk
¹Õ ÁÑ¡¨ÐÁÕÅѡɳÐà» ¹ÊÑҳú¡Ç¹áººÊÕ (colored noise) ·Õ ÁÕ¤èÒà©ÅÕ Âà·èÒ
¡Ñº¤èÒÈÙ¹ÂìáÅФèÒ¤ÇÒÁá»Ã»Ãǹà·èҡѺ
2 σw
¨Ò¡áºº¨ÓÅͧã¹ÃÙ» ·Õ 5.2 ÊÑÒ³ÍÔ¹¾Øµ ·Õ ´éÒ¹¢Òà¢éÒ Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÊÒÁÒöà¢Õ¹ãËé ÍÂÙè
94
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ã¹ÃÙ»ÊÁ¡Ò÷ҧ¤³ÔµÈÒʵÃìã¹â´àÁ¹
D
ä´é´Ñ§¹Õ
Y (D) = A(D)H(D) + W (D) = X(D) + W (D) â´Â·Õ
A(D) ¤×Í ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ, H(D) =
¢Í§·ÒÃìà¡çµ,
X(D)
=
A(D)H(D)
PL−1
hk Dk
k=0
(5.1)
¤×Í ·ÒÃìà¡çµ,
L ¤×Í ¨Ó¹Ç¹á·ç»·Ñ §ËÁ´
¤×Í ¢éÍÁÙÅ àÍÒµì¾Øµ ªèͧÊÑÒ³, áÅÐ
ú¡Ç¹áººÊÕ·Õ ÁÕ¤èÒà©ÅÕ Âà·èҡѺ¤èÒÈÙ¹Âì áÅФèÒ¤ÇÒÁá»Ã»Ãǹà·èҡѺ ãËé ¾Ô¨ÒóÒÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ 2 Ẻ ¤×Í
A1 (D)
áÅÐ
W (D)
¤×Í ÊÑÒ³
2 σw
A2 (D)
«Ö § à» ¹ ÅӴѺ ¢éÍÁÙÅ ·Õ ·ÓãËé
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ à¡Ô´ ¤ÇÒÁÊѺʹã¹ÃÐËÇèÒ§¡Ãкǹ¡ÒöʹÃËÑÊ ÁÒ¡·Õ ÊØ´ â´Â·Õ ÊÁÒªÔ¡ áµèÅÐ µÑÇã¹ÅӴѺ¢éÍÁÙÅÍÔ¹¾ØµÁÕ¤èÒ·Õ à» ¹ä»ä´é ¤×Í
{−1, 1}
´Ñ§¹Ñ ¹ ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ ÊÍ´¤Åéͧ¡Ñº
à˵ءÒóì¢éͼԴ¾ÅÒ´ ¨Ð¹ÔÂÒÁâ´Â
εa (D) = A1 (D) − A2 (D) àÁ× Í ÊÁÒªÔ¡áµèÅеÑÇã¹
εa (D)
à» ¹ÊÁÒªÔ¡¢Í§
{−2,
(5.2)
0, 2} ¶éÒ¡Ó˹´ãËéÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾ØµÁÕ
ÃÙ»¢Í§¾ËعÒÁ (polynomial) ¤×Í
εa (D) =
p−1 X
εa,k Dk
k=0
= εa,0 + εa,1 D + · · · + εa,p−1 Dp−1
(5.3)
ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ ¶Ù¡µéͧ (valid input error sequence) ¨ÐµéͧÊÍ´¤Åéͧ¡Ñº ¤Ø³ÊÁºÑµÔ 2 ¢éÍ ´Ñ§µèÍ仹Õ
1) ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¾¨¹ìáááÅо¨¹ìÊØ´·éÒ¢ͧ
0
áÅÐ
εa (D)
µéͧÁÕ¤èÒäÁèà·èҡѺ¤èÒÈÙ¹Âì ¹Ñ ¹¤×Í
εa,0 6=
εa,p−1 6= 0
2) ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ
εa (D)
·Õ ¶Ù¡µéͧ¨ÐµéͧäÁè Á¤ Õ èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ·Õ ÁÕ ¤èÒ à·èÒ ¡Ñº ¤èÒ ÈÙ¹Âì ã¹
ÅӴѺ¢éͼԴ¾ÅÒ´àÃÕ§µÔ´¡Ñ¹à» ¹¨Ó¹Ç¹ÁÒ¡¡ÇèÒËÃ×Íà·èҡѺ ä´é¨ÐÊ觼ŷÓãËéà¡Ô´à˵ءÒóì¢éͼԴ¾ÅÒ´·Õ àËÁ×͹¡Ñ¹ [49]
L−1
µÑÇ ÁÔ©Ð¹Ñ ¹áÅéÇ ¼ÅÅѾ¸ì·Õ
5.2.
¤ÇÒÁËÁÒ¢ͧà˵ءÒóì¢éͼԴ¾ÅÒ´
95
{ 1 1 -1 1 -1 1 1 -1 -1 1 1 -1 1 -1 1 1 -1 1 -1 1 } A2 ( D ) = { 1 -1 1 -1 -1 1 -1 1 -1 -1 1 -1 -1 1 1 1 1 -1 1 1 } A1 ( D ) =
ε a ( D) =
{ 0 2 -2 2 0 0 2 -2 0 2 0 0 2 -2 0 0 -2 2 -2 0 } ÃÙ»·Õ 5.3: µÑÇÍÂèÒ§¡ÒäӹdzËÒÅӴѺ¢éͼԴ¾ÅÒ´
àÁ× Í ä´é ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Õ ¶Ù¡µéͧáÅéÇ à˵ءÒÃ³ì ¢éͼԴ¾ÅÒ´ (error event) ¢Í§Ãкº¨Ð ¹ÔÂÒÁâ´Â
εx (D) = εa (D)H(D)
µÑÇÍÂèÒ§·Õ 5.1
¾Ô¨ÒóÒẺ¨ÓÅͧªèͧÊÑÒ³ PR4,
H(D) = 1 − D2 ,
ÍÔ¹¾ØµÁÕ¤ÇÒÁÂÒÇà·èҡѺ 20 ºÔµ â´Â·Õ ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ·Õ Ë¹Ö §
−1, −1, −1,
1,
1, 1,
−1,
−1, −1,
1,
1,
−1,
−1,
1,
1, 1,
−1,
−1, −1,
−1, 1}
1,
1,
−1, −1,
(5.4)
A1 (D)
¶éÒ¡Ó˹´ãËéÅӴѺ¢éÍÁÙÅ
¤×Í
{1,
1, 1, 1,
áÅÐÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ·Õ Êͧ 1, 1, 1, 1,
−1,
−1,
A2 (D)
1, 1,
¤×Í
{1,
1, 1} ¨§ËÒÅӴѺ¢éͼԴ¾ÅÒ´
ÍÔ¹¾Øµ·Õ ¶Ù¡µéͧ·Ñ §ËÁ´·Õ ¾ºã¹Ãкº¹Õ
ÇÔ¸Õ ·Ó
àÃÔ Áµé¹ ãËé ¤Ó¹Ç³ËÒÅӴѺ ¢éͼԴ¾ÅÒ´
εa (D)
¨Ò¡ÊÁ¡Òà (5.2) «Ö § ¨Ðä´é ¼ÅÅѾ¸ì ´Ñ§ áÊ´§
ã¹ÃÙ» ·Õ 5.3 ¨Ò¡¹Ñ ¹ ãËé ·Ó¡ÒÃËÒÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Õ ¶Ù¡µéͧ «Ö § ¨Ò¡ÃÙ» ·Õ 5.3 ¨Ðä´é ÇèÒ ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ ¶Ù¡µéͧÁÕ·Ñ §ËÁ´ 4 µÑÇ ¤×Í
{2, −2, 2},
{2, −2},
{2, −2, 2}
áÅÐ
{−2,
A1 (D) − A2 (D)
ËÃ×Í
A2 (D) − A1 (D)
áµè ÅӴѺ ¢éͼԴ¾ÅÒ´ ¨Ð¾Ô¨ÒóҨҡ
{2, −2, 0, 2}, 2,
−2}
{−2, 2, −2}
¶×Í ÇèÒ à» ¹ µÑÇ à´ÕÂǡѹ à¹× ͧ¨Ò¡¢Ö ¹ ÍÂÙè ¡Ñº ÇèÒ à¾ÃÒÐ©Ð¹Ñ ¹ ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ
¶Ù¡µéͧã¹Ãкº¨Ð¶×ÍÇèÒÁÕ·Ñ §ËÁ´ 3 Ẻ ¤×Í
{2, −2, 2},
{2, −2, 0, 2},
{2, −2}
96
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ËÃ×ÍÊÒÁÒöà¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§ÊÁ¡Òä³ÔµÈÒʵÃìã¹â´àÁ¹
2D3 ,
áÅÐ
ËÁÒÂà˵Ø
2 − 2D
ä´é ¤×Í
2 − 2D + 2D2 , 2 − 2D +
µÒÁÅӴѺ
¨ÐàËç¹ä´éÇèÒ ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ ¶Ù¡µéͧ
¢Ö ¹ÍÂÙè¡ÑºÇèҨоԨÒóҨҡ
µÑÇÍÂèÒ§·Õ 5.2
D
A1 (D) − A2 (D)
ËÃ×Í
εa (D)
¨ÐÁÕ¼ÅÅѾ¸ìà·èҡѺ
−εa (D)
·Ñ §¹Õ
{−1, 1}
à» ¹
A2 (D) − A1 (D)
¡Ó˹´ãËéÊÁÒªÔ¡áµèÅеÑǢͧÅӴѺ¢éÍÁÙÅÍÔ¹¾ØµÁÕ¤èÒ·Õ à» ¹ä»ä´é ¤×Í
·Õ ·ÃÒº¡Ñ¹ÇèÒ ã¹Ãкº¡Òúѹ·Ö¡áººá¹Ç¹Í¹ (longitudinal recording) ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ à¡Ô´¢Ö ¹ºèÍÂã¹ÃÐËÇèÒ§¡Ãкǹ¡ÒöʹÃËÑÊ¢éÍÁÙÅ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¤×Í ¤Ó¹Ç³ËÒà˵ءÒóì¢éͼԴ¾ÅÒ´¢Í§Ãкº¹Õ àÁ× ÍÃкºãªé·ÒÃìà¡çµáºº PR4,
ÇÔ¸Õ·Ó
¨Ò¡·Õ ⨷Âì¡Ó˹´ãËé ¨Ðä´éÇèÒ
εa (D) = 2 − 2D + 2D2
{2, −2, 2}
[19] ¨§
H(D) = 1 − D2
´Ñ§¹Ñ ¹ à˵ءÒóì¢éͼԴ¾ÅÒ´
εx (D)
ÊÒÁÒöËÒä´é¨Ò¡ÊÁ¡Òà (5.4) ¹Ñ ¹¤×Í
εx (D) = εa (D)H(D) = (2 − 2D + 2D2 )(1 − D2 ) = 2 − 2D + 2D3 − 2D4 à¾ÃÒÐ©Ð¹Ñ ¹ à˵ءÒÃ³ì ¢éͼԴ¾ÅÒ´¢Í§Ãкº¡Òúѹ·Ö¡ Ẻá¹Ç¹Í¹ ·Õ ãªé ·ÒÃìà¡çµ Ẻ PR4 ¤×Í
εx (D) = 2 − 2D + 2D3 − 2D4
µÑÇÍÂèÒ§·Õ 5.3
¡Ó˹´ãËéÊÁÒªÔ¡áµèÅеÑǢͧÅӴѺ¢éÍÁÙÅÍÔ¹¾ØµÁÕ¤èÒ·Õ à» ¹ä»ä´é ¤×Í
{−1, 1}
à» ¹
·Õ ·ÃÒº¡Ñ¹ ÇèÒ ã¹Ãкº¡Òúѹ·Ö¡ Ẻá¹ÇµÑ § (perpendicular recording) ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Õ à¡Ô´¢Ö ¹ºèÍÂã¹ÃÐËÇèÒ§¡Ãкǹ¡ÒöʹÃËÑÊ¢éÍÁÙÅ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¤×Í ¤Ó¹Ç³ËÒà˵ءÒóì¢éͼԴ¾ÅÒ´¢Í§Ãкº¹Õ àÁ× ÍÃкºãªé·ÒÃìà¡çµáºº PR2,
ÇÔ¸Õ ·Ó
¨Ò¡·Õ ⨷Âì ¡Ó˹´ãËé ¨Ðä´é ÇèÒ
εa (D) = 2 − 2D
{2, −2}
[18] ¨§
H(D) = 1 + 2D + D2
´Ñ§¹Ñ ¹ à˵ءÒÃ³ì ¢éͼԴ¾ÅÒ´
εx (D)
5.3.
ÃÐÂзҧÂؤÅÔ´
97
ÊÒÁÒöËÒä´é¨Ò¡ÊÁ¡Òà (5.4) ¹Ñ ¹¤×Í
εx (D) = εa (D)H(D) = (2 − 2D)(1 + 2D + D2 ) = 2 + 2D − 2D2 − 2D3 à¾ÃÒÐ©Ð¹Ñ ¹ à˵ءÒóì¢éͼԴ¾ÅÒ´¢Í§Ãкº¡Òúѹ·Ö¡áººá¹ÇµÑ § ·Õ ãªé·ÒÃìà¡çµáºº PR2 ¤×Í =
εx (D)
2 − 2D + 2D3 − 2D4
5.3
ÃÐÂзҧÂؤÅÔ´
¨Ò¡ÃÙ» ·Õ 5.2 ÅӴѺ ¢éͼԴ¾ÅÒ´¢Í§ÊÑÒ³·Õ ´éÒ¹¢Òà¢éÒ ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ËÃ×Í ·Õ àÃÕ¡¡Ñ¹ ÇèÒ à˵ءÒóì¢éͼԴ¾ÅÒ´ (error event) ¢Í§Ãкº ¤×Í
εx (D) = [A1 (D) − A2 (D)] H(D) = X1 (D) − X2 (D)
(5.5)
¡Ãкǹ¡ÒöʹÃËÑÊ ¢éÍÁÙÅ â´Âãªé ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÊÒÁÒö·Õ ¨Ð¶Ù¡ ͸ԺÒÂâ´Âãªé ÃÙ»ÀÒ¾ÊͧÁÔµÔ µÒÁÃÙ»·Õ 5.4 àÁ× Í
X1 (D)
¤×Í ÊÑÒ³·Õ ¶Ù¡µééͧ áÅÐ
X2 (D)
ÇÕà·ÍÃìºÔ¨ÐµÑ´ÊԹ㨼Դ¾ÅÒ´ ¶éÒ¢¹Ò´¢Í§ÊÑҳú¡Ç¹ ä»Âѧ
X2 (D)
ÃÐËÇèÒ§
ÁÕ¤èÒÁÒ¡¡ÇèÒ
X1 (D)
áÅÐ
d/2
â´Â·Õ
d
¤×Í ÊÑÒ³·Õ ¼Ô´¾ÅÒ´ ǧ¨ÃµÃǨËÒ
W (D) (¹Ñ ¹¤×Í ϕ) ã¹·ÔÈ·Ò§¨Ò¡ X1 (D)
¤×Í ÃÐÂзҧÂؤÅÔ´ (Euclidean distance)
εx (D)
·Õ ÇÑ´
X2 (D)
¶éÒ¡Ó˹´ãËéÅӴѺà˵ءÒóì¢éͼԴ¾ÅÒ´¢Í§ÃкºÁÕÃÙ»¢Í§¾ËعÒÁ ¤×Í
εx (D) =
n X
εx,k Dk
k=0
= εx,0 + εx,1 D + · · · + εx,n Dn ´Ñ§¹Ñ ¹ÃÐÂзҧÂؤÅÔ´¡ÓÅѧÊͧ (squared Euclidean distance) ¢Í§à˵ءÒóì¢éͼԴ¾ÅÒ´,
(5.6)
d2 {εa (D)},
98
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
Y ( D) W ( D)
ϕ
X 1 ( D)
X 2 ( D)
d Decision boundary ÃÙ»·Õ 5.4: ÀÒ¾ÊͧÁÔµÔáÊ´§¡Ãкǹ¡ÒöʹÃËÑÊ¢éÍÁÙÅâ´Âãªéǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
¨Ð¶Ù¡¹ÔÂÒÁãËéÁÕ¤èÒà·èҡѺ ¤èÒ¾Åѧ§Ò¹¢Í§
εx (D)
[15] ¹Ñ ¹¤×Í
d2 {εa (D)} = kεx (D)k2 n X ε2x,k = k=0
= ε Tε â´Â·Õ
k·k
·Õ ÁÕÊÁÒªÔ¡
¤×Í ¡ÒÃËÒ¤èÒ¹ÍÃìÁ (norm) áÅÐ
ε
¤×Í àÇ¡àµÍÃìá¹ÇµÑ §¢Í§à˵ءÒóì¢éͼԴ¾ÅÒ´
n+1 µÑÇ µÑÇÍÂèÒ§àªè¹ ¶éÒ εx (D) = 2−3D+4D3 −5D4
µÑÇÍÂèÒ§·Õ 5.4
(5.7)
¨Ðä´éÇèÒ
εx (D)
ε = [2, −3, 0, 4, −5]T
¨Ò¡¢éÍÁÙÅ ã¹µÑÇÍÂèÒ§·Õ 5.3 ¨§¤Ó¹Ç³ËÒÃÐÂзҧÂؤÅÔ´ ¡ÓÅѧ Êͧ¢Í§à˵ءÒóì
¢éͼԴ¾ÅÒ´¢Í§Ãкº
ÇÔ¸Õ·Ó
¨Ò¡µÑÇÍÂèÒ§·Õ 5.3 ¨Ðä´éÇèÒ
2D + 2D3 − 2D4
εa (D) = 2 − 2D, H(D) = 1 + 2D + D2 ,
´Ñ§¹Ñ ¹ ÃÐÂзҧÂؤÅÔ´¡ÓÅѧÊͧ¢Í§à˵ءÒóì¢éͼԴ¾ÅÒ´
áÅÐ
εx (D) = 2 −
d2 {εa (D)}
ËÒä´é¨Ò¡ÊÁ¡Òà (5.7) ´Ñ§µèÍ仹Õ
2
d {εa (D)} =
n X i=0
ε2x,i = (2)2 + (−2)2 + (0)2 + (2)2 + (−2)2 = 16
ÊÒÁÒö
5.4.
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å
99
ËÃ×ÍÍÒ¨¨ÐËÒä´é¨Ò¡
d2 {εa (D)} = ε Tε = [2, −2, 0, 2, −2] · [2, −2, 0, 2, −2]T = 16 à¾ÃÒÐ©Ð¹Ñ ¹ ÃÐÂзҧÂؤÅÔ´¡ÓÅѧÊͧ¢Í§à˵ءÒóì¢éͼԴ¾ÅÒ´
εx (D) = 2 − 2D + 2D3 − 2D4
ÁÕ¤èÒà·èҡѺ 16
¶éÒÊÁÁصÔÇèÒ ÊÑҳú¡Ç¹ ¤èÒ¤ÇÒÁá»Ã»Ãǹ¢Í§
W (D)
W (D)
à» ¹ÊÑҳú¡Ç¹à¡ÒÊìÊÕ¢ÒÇẺºÇ¡ (AWGN) ¨Ðä´éÇèÒ
¨ÐÁÕ¢¹Ò´à·èҡѹ㹷ء·ÔÈ·Ò§ ´Ñ§¹Ñ ¹ »ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ç§¨ÃµÃǨËÒ
ÇÕà·ÍÃìºÔã¹ÃÙ»¢Í§ÍѵÃÒ¢éͼԴ¾ÅÒ´ºÔµ (BER) ¨Ð¢Ö ¹ÍÂÙè¡ÑºÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·ÓãËéà¡Ô´ÅӴѺ¢éͼԴ¾ÅÒ´
εx (D)
εa (D)
·Õ à» ¹¼Å
·Õ ÁÕ¤èÒÃÐÂзҧÂؤÅÔ´¡ÓÅѧÊͧ·Õ ¹éÍÂÊØ´ [15] «Ö §¹ÔÂÒÁâ´Â
d2min =
min
valid εa (D)
£ 2 ¤ d {εa (D)}
(5.8)
à¾ÃÒÐ©Ð¹Ñ ¹ ¤ÇÒÁ¹èÒ¨Ðà» ¹ ¢Í§¢éͼԴ¾ÅÒ´ (probability of error) ËÃ×Í BER ·Õ ´éÒ¹¢ÒÍÍ¡¢Í§ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÊÒÁÒö»ÃÐÁÒ³¤èÒä´é´Ñ§¹Õ [15]
µ Pe ≈ K1 Q àÁ× Í
K1
¤×Í ¤èÒ¤§µÑÇ·Õ äÁè¢Ö ¹¡Ñº¤èÒ
σw
áÅÐ
Q(x) =
dmin 2σw
√1 2π
¶
R∞ x
(5.9)
e−
u2 2
du ´Ñ§¹Ñ ¹ ã¹¡Ã³Õ·Õ Í§¤ì»ÃСͺ
¢Í§ÊÑҳú¡Ç¹·Õ ´éÒ¹¢Òà¢éҢͧǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔà» ¹ÊÑҳú¡Ç¹ AWGN ¤èÒ ÃÐÂзҧ ÂؤÅÔ´·Õ ¹éÍÂÊØ´
dmin
ÊÒÁÒö·Õ ¨Ð¹ÓÁÒãªé㹡ÒûÃÐÁÒ³¤èÒ¤ÇÒÁ¹èÒ¨Ðà» ¹¢Í§¢éͼԴ¾ÅÒ´¢Í§Ãкº
ä´é â´Â·Õ ¼ÅÅѾ¸ì ·Õ ä´é ¨ÐÁÕ ¤ÇÒÁ¹èÒ àª× Ͷ×Í ÁÒ¡ÂÔ § ¢Ö ¹ ¶éÒ ÃкºÁÕ à˵ءÒÃ³ì ¢éͼԴ¾ÅÒ´·Õ â´´à´è¹ (dominant error event) à¾Õ§µÑÇà´ÕÂÇ ¹Ñ ¹¤×ÍÁÕ¤èÒ
d2 {εa (D)} ¹éÍÂÁÒ¡ àÁ× Íà·Õº¡Ñº¤èÒ d2 {εa (D)}
¢Í§ÅӴѺ¢éͼԴ¾ÅÒ´Í× ¹æ ·Õ ¾ºã¹Ãкº
5.4
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å
¶éÒÊÑҳú¡Ç¹·Õ ´éÒ¹¢Òà¢éҢͧǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
W (D)
à» ¹ÊÑҳú¡Ç¹áººÊÕ («Ö §ã¹
·Ò§»¯ÔºÑµÔáÅéÇÁÑ¡¨Ðà» ¹àªè¹¹Õ â´Â੾ÒÐÍÂèÒ§ÂÔ §·Õ ND ÊÙ§) ¤ÇÒÁ¹èÒ¨Ðà» ¹¢Í§¢éͼԴ¾ÅÒ´¨Ð¢Ö ¹ÍÂÙè
100
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¡Ñº ·Ñ § ÃÐÂзҧÂؤÅÔ´ áÅФèÒ ¤ÇÒÁá»Ã»Ãǹ¢Í§ÊÑҳú¡Ç¹ã¹·ÔÈ·Ò§¢Í§ ã¹¡Ã³Õ¹Õ ¤èÒ
dmin
dmin
X1 (D)
ä»Âѧ
X2 (D)
ε
W (D)
àÇ¡àµÍÃì á¹ÇµÑ § ¢Í§ÊÑҳú¡Ç¹ ¨Ðä´éÇèÒ
W (D)
(5.10)
εx (D)
·Õ ÁÕ ÊÁÒªÔ¡
·Õ ÁÕ ÊÁÒªÔ¡
n+1
n+1
µÑÇ àªè¹ ¶éÒ
µÑÇ, áÅÐ
w
W (D) = 0.82 −
ϕ
ÊÒÁÒöËÒä´é¨Ò¡ [48]
σϕ2 = E[ϕ2 ] · T T T ¸ (w ε ) (w ε ) = E kεεk2 £ ¤ ε T E wwT ε = ε Tε T ε Rwwε = ε Tε E[·]
¤×Í
w = [0.82, −1.3, 0, 0.2]T
´Ñ§¹Ñ ¹ ¤èÒ¤ÇÒÁá»Ã»Ãǹ¢Í§
àÁ× Í
ϕ
ŧº¹àÇ¡àµÍÃìË¹Ö §Ë¹èÇ (unit vector)
w Tε kεεk
¤×Í àÇ¡àµÍÃì á¹ÇµÑ § ¢Í§à˵ءÒÃ³ì ¢éͼԴ¾ÅÒ´
1.3D + 0.2D3
[48, 49]
¹Ñ ¹¤×Í
ϕ= àÁ× Í
deff {εa (D)}
ÊÓËÃѺ¡ÒûÃÐÁÒ³¤èÒ¤ÇÒÁ¹èÒ¨Ðà» ¹¢Í§¢éͼԴ¾ÅÒ´¢Í§Ãкº ¨Ò¡ÃÙ»·Õ 5.4 ¤èÒ
ÁÕ¤èÒà·èҡѺ¡ÒéÒ (projection) ÊÑҳú¡Ç¹ ·Õ ÁÕ·ÔÈ·Ò§¨Ò¡
´Ñ§¹Ñ ¹
¨Ö§äÁèÊÒÁÒö¹ÓÁÒãªé㹡ÒûÃÐÁÒ³¤èÒ¤ÇÒÁ¹èÒ¨Ðà» ¹¢Í§¢éͼԴ¾ÅÒ´·Õ à¡Ô´¢Ö ¹ã¹
Ãкºä´é ´Ñ§¹Ñ ¹ ¨Ö§ä´éÁÕ¡ÒÃ¹Ó ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å (e ective distance) ÁÒãªéá·¹
εx (D), ϕ,
¤×Í µÑÇ´Óà¹Ô¹¡ÒäèÒ¤Ò´ËÁÒ (expectation operator),
(auto correlation matrix) ¢Í§
W (D)
ËÒä´é¨Ò¡
Rww (i, j) = E
â´Â·Õ ÊÁÒªÔ¡á¶Ç·Õ
"S−1 X
i
(5.11)
Rww
¤×Í àÁ·ÃÔ¡«ìÍѵÊËÊÑÁ¾Ñ¹¸ì
áÅÐá¹ÇµÑ §·Õ
j
¢Í§àÁ·ÃÔ¡«ì
Rww
# wk−i wk−j ,
0 ≤ i, j ≤ n + 1
(5.12)
k=0 àÁ× Í
S
¤×Í ¤ÇÒÁÂÒǢͧÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ
SNRevent {εa (D)},
{ak }
ã¹ [49] ¤èÒ SNR ¢Í§à˵ءÒÃ³ì ¢éͼԴ¾ÅÒ´,
¨Ð¹ÔÂÒÁâ´Â
SNRevent {εa (D)} =
d2 {εa (D)} 2 σw
(5.13)
5.4.
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å
áÅÐ SNR »ÃÐÊÔ·¸Ô¼Å,
101
SNReff ,
¢Í§à˵ءÒóì¢éͼԴ¾ÅÒ´ ¤×Í
d2 {εa (D)} σϕ2
SNReff {εa (D)} =
(5.14)
à¾× ÍãËéÊÒÁÒö»ÃÐÁÒ³¤èÒ¤ÇÒÁ¹èÒ¨Ðà» ¹¢Í§¢éͼԴ¾ÅÒ´·Õ à¡Ô´¢Ö ¹ã¹Ãкºä´éÍÂèÒ§¹èÒàª× Ͷ×ÍÁÒ¡ ¡ÇèÒ ¡ÒÃãªé ¾ÒÃÒÁÔàµÍÃì
dmin
µÒÁÊÁ¡Òà (5.9) ¨Ö§ ä´é ÁÕ ¡ÒùÔÂÒÁ¤èÒ ÃÐÂзҧẺãËÁè ¢Ö ¹ ÁÒ·Õ àÃÕ¡
ÇèÒ ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å (e ective distance), á»Ã»Ãǹ
σϕ2
à¢éÒä»´éÇÂáÅéÇ ´Ñ§¹Ñ ¹ àÁ× Íá·¹¤èÒ
áÅéÇ ¨Ðà» ¹¼Å·ÓãËé
SNRevent {εa (D)}
deff {εa (D)}, d2 {εa (D)}
ÁÕ¤èÒà·èҡѺ
SNReff {εa (D)} =
«Ö § ¨ÐÃÇÁ¼Å¡Ãзº¢Í§¤èÒ ¤ÇÒÁ
d2eff {εa (D)}
´éÇÂ
SNReff {εa (D)}
ã¹ÊÁ¡Òà (5.13)
¹Ñ ¹¤×Í
d2eff {εa (D)} d2 {εa (D)} = 2 σw σϕ2
(5.15)
¨Ò¡ÊÁ¡Òà (5.15) ¨Ðä´éÇèÒ ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧÊͧ (squared e ective distance) ÁÕ¤èÒà·èҡѺ
2 d2eff {εa (D)} = σw 2 = σw
d2 {εa (D)} σϕ2 (εεTε )2 ε T Rwwε
(5.16)
àªè¹à´ÕÂǡѹ ¤èÒÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧÊͧ·Õ ¹éÍÂÊØ´ ¨Ð¹ÔÂÒÁâ´Â
d2effmin =
£
min
valid εa (D)
¤ d2eff {εa (D)}
(5.17)
áÅФÇÒÁ¹èÒ¨Ðà» ¹¢Í§¢éͼԴ¾ÅÒ´ã¹ (5.9) ¨Ðà»ÅÕ Â¹ä»à» ¹ [15]
µ Pe ≈ K2 Q àÁ× Í
K2
¤×Í ¤èÒ¤§µÑÇ·Õ äÁè¢Ö ¹¡Ñº¤èÒ
σw
deffmin 2σw
¶
¨Ò¡¡Ò÷´Åͧ¾ºÇèÒ
(5.18)
deffmin
ÊÒÁÒö¹Óãªé㹡ÒûÃÐÁÒ³
¤èÒ ¤ÇÒÁ¹èÒ¨Ðà» ¹ ¢Í§¢éͼԴ¾ÅÒ´¢Í§Ãкºä´é ÍÂèÒ§¹èÒ àª× Ͷ×Í ÁÒ¡¡ÇèÒ ¡ÒÃãªé
deffmin
dmin
·Ñ §¹Õ à¹× ͧÁÒ¨Ò¡
ä´éÃÇÁ¼Å¡Ãзº·Õ à¡Ô´¨Ò¡ÊÑҳú¡Ç¹áººÊÕäÇéáÅéÇ
µÑÇÍÂèÒ§·Õ 5.5
¾Ô¨ÒóÒẺ¨ÓÅͧ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ ã¹ÃÙ» ·Õ 3.2 ÊÓËÃѺ Ãкº¡Òúѹ·Ö¡ Ẻ
á¹ÇµÑ § ·Õ ND = 2 áÅÐ SNR = 22 dB â´Â·Õ ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
ak ∈ {−1, 1}
áÅлÃÒ¡®ÇèÒ·Ó
102
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ãËéä´é ÅӴѺ¢éͼԴ¾ÅÒ´
−12.72}
{wk }
=
{−3.64, −4.34,
¶éÒÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ
Ẻ PR2,
H(D) = 1 + 2D + D2
εa (D)
2.96,
−1.56, −3.70,
·Õ à¡Ô´¢Ö ¹ºèÍÂÃкº ¤×Í
0.80, 8.52,
{2, −2}
−3.76,
6.10,
áÅÐÃкºãªé·ÒÃìà¡çµ
¨§¤Ó¹Ç³ËÒÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧÊͧ
d2eff {εa (D)}
¢Í§
ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ
ÇÔ¸Õ·Ó
¨Ò¡·Õ ⨷Âì¡Ó˹´ à˵ءÒóì¢éͼԴ¾ÅÒ´
εx (D)
¢Í§Ãкº¹Õ ËÒä´é¨Ò¡
εx (D) = εa (D)H(D) = (2 − 2D)(1 + 2D + D2 ) = 2 + 2D − 2D2 − 2D3 ¹Ñ ¹¤×Í
ε = [2, 2, −2, −2]T
¨Ò¡ÅӴѺ ¢éͼԴ¾ÅÒ´
{wk }
·Õ ¡Ó˹´ãËé ¨Ðä´é ÇèÒ àÁ·ÃÔ¡«ìÍѵÊËÊÑÁ¾Ñ¹¸ì
Rww
ÊÒÁÒöËÒä´é
¨Ò¡ÊÁ¡Òà (5.12) «Ö §ÁÕ¤èÒà·èҡѺ
34.33 −13.84
Rww
áÅÐ
6.13 −11.25
−13.84 34.33 −13.84 6.13 = 6.13 −13.84 34.33 −13.84 −11.25 6.13 −13.84 34.33
2 = R σw ww (0, 0) = 34.33
´Ñ§¹Ñ ¹ ¤èÒ ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧ Êͧ
d2eff {εa (D)}
ËÒä´é ¨Ò¡
ÊÁ¡Òà (5.16) ¹Ñ ¹¤×Í
(εεTε )2 ε T Rwwε µ ¶ 256 = 34.33 430.48 = 20.41
2 d2eff {εa (D)} = σw
à¾ÃÒÐ©Ð¹Ñ ¹ ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ
(5.19)
εa (D) = {2, −2} ã¹Ãкº¡Òúѹ·Ö¡áººá¹ÇµÑ §·Õ ãªé·ÒÃìà¡çµ
Ẻ PR2 ¨ÐÁÕ¤èÒÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧÊͧà·èҡѺ 20.41 ˹èÇÂ
5.5.
¼Å¡Ò÷´Åͧ
103
ËÅѧ¨Ò¡¤Ó¹Ç³ËÒ¤èÒ
deffmin
deffmin
ä´é áÅéÇ ´Ñ§¹Ñ ¹ ¤èÒ SNR »ÃÐÊÔ·¸Ô¼Å
¨Ð¹ÔÂÒÁâ´Â
SNReff =
SNReff
d2effmin (εεTε )2 = 2 σw ε T Rwwε
¨Ò¡¡Ò÷´Åͧ·Õ ¨ÐáÊ´§ã¹ËÑÇ¢éÍ·Õ 5.5 ¨Ð¾ºÇèÒ
SNReff
·Õ ÊÍ´¤Åéͧ¡Ñº
(5.20)
ÊÒÁÒö¹ÓÁÒãªé㹡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸Ô
ÀÒ¾¢Í§Ãкº·Õ ãªé ·ÒÃìà¡çµ ẺµèÒ§æ ä´é ÍÂèÒ§¹èÒ àª× Ͷ×Í àªè¹à´ÕÂǡѹ ¡Ñº ¡ÒÃãªé ¾ÒÃÒÁÔàµÍÃì BER ã¹ ¡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº â´Â¼ÅÅѾ¸ì ·Õ ä´é ¨ÐÁÕ ¤ÇÒÁ¹èÒ àª× Ͷ×Í ÁÒ¡ÂÔ § ¢Ö ¹ ¶éÒ ã¹ÃкºÁÕ à˵ءÒÃ³ì ¢éͼԴ¾ÅÒ´·Õ â´´à´è¹ à¾Õ§µÑÇ à´ÕÂÇà·èÒ¹Ñ ¹ ¹Ñ ¹¤×Í àÁ× Í Ãкº·Ó§Ò¹ ³ ÃдѺ SNR ·Õ ÊÙ§ à¾Õ§¾Í àªè¹ àÁ× ÍÃкºÁÕ BER
5.5
< 10−4
¼Å¡Ò÷´Åͧ
㹡Ò÷´Åͧà¾× Í ·Ó¡ÒÃÇÔà¤ÃÒÐËì à˵ءÒÃ³ì ¢éͼԴ¾ÅÒ´ ¨Ð¾Ô¨ÒóÒ੾ÒÐÃкº¡Òúѹ·Ö¡ Ẻá¹Ç µÑ § (perepdicular recording) à·èÒ¹Ñ ¹ ÊÓËÃѺ¡ÒÃÇÔà¤ÃÒÐËìà˵ءÒóì¢éͼԴ¾ÅÒ´¢Í§Ãкº¡Òúѹ·Ö¡ Ẻá¹Ç¹Í¹ (longitudinal recording) ÊÒÁÒöÈÖ¡ÉÒä´é¨Ò¡ [49] ãËé¾Ô¨ÒóÒẺ¨ÓÅͧÃкºã¹ ÃÙ»·Õ 3.2 â´Â·Õ ÊÑÒ³ read back ¨Ðà» ¹ä»µÒÁÊÁ¡Òà (3.26) ¹Ñ ¹¤×Í
p(t) =
S−1 X
bk g(t − kT + ∆tk ) + n(t)
(5.21)
k=0 àÁ× Í
bk = (ak − ak−1 )/2
ʶҹкǡËÃ×Íź áÅÐ ·Õ
k
·Õ Áըӹǹ·Ñ §ËÁ´
¤×Í ºÔµ à»ÅÕ Â¹Ê¶Ò¹Ð (àÁ× Í
bk = 0
bk = ±1
¤×ÍäÁèÁÕ¡ÒÃà»ÅÕ Â¹á»Å§Ê¶Ò¹Ð),
S = 4096
ºÔµ (1 à«¡àµÍÃì),
¡Òúѹ·Ö¡ Ẻá¹ÇµÑ § µÒÁÊÁ¡Òà (1.2),
n(t)
g(t)
ÊÍ´¤Åéͧ¡Ñº ¡ÒÃà»ÅÕ Â¹á»Å§
ak ∈ ±1
¤×Í ¢éÍÁÙÅÍÔ¹¾ØµºÔµµÑÇ
¤×Í ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹Ð¢Í§Ãкº
¤×Í ÊÑҳú¡Ç¹à¡ÒÊì ÊÕ ¢ÒÇẺºÇ¡ (AWGN)
·Õ ÁÕ¤ÇÒÁ˹Òá¹è¹Ê໡µÃÑÁ¡ÓÅѧẺÊͧ´éÒ¹à·èҡѺ
N0 /2,
áÅÐ
∆t
¤×Í ÊÑҳú¡Ç¹¨ÔµàµÍÃì
¢Í§Ê× ÍºÑ¹·Ö¡ (media jitter noise) ·Õ ¶Ù¡¨ÓÅͧãËéÁÕÅѡɳÐà» ¹ ¡ÒÃàÅ× Í¹µÓá˹觢ͧ¡ÒÃà»ÅÕ Â¹ ʶҹÐẺÊØèÁ «Ö §ÁÕ¿ §¡ìªÑ¹¤ÇÒÁ˹Òá¹è¹¤ÇÒÁ¹èÒ¨Ðà» ¹áººà¡ÒÊìà«Õ¹·Õ ÁÕ¤èÒà©ÅÕ Âà·èҡѺ¤èÒÈÙ¹Âì áÅФèÒ ¤ÇÒÁá»Ã»Ãǹà·èÒ ¡Ñº
|bk |σj2
áÅж١ ¨Ó¡Ñ´ ãËé ÁÕ ¤èÒ äÁè à¡Ô¹
¨Ó¹Ç¹à»ÍÃìà«ç¹µì¢Í§ºÔµà«ÅÅì
T
|bk |
áÅÐ
¤×Í ¤èÒÊÑÁºÙóì¢Í§
bk
T /2
àÁ× Í
σj
¨Ð¶Ù¡ ¡Ó˹´à» ¹
104
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ÊÑÒ³ read back ¨Ð¶Ù¡ Êè§ ¼èÒ¹ä»Âѧ ǧ¨Ã¡Ãͧ¼èÒ¹µ Ó ºÑµà·ÍÃìàÇÔÃìµ Íѹ´Ñº ·Õ 7 áÅж١ ·Ó
1/T
¡Òêѡ µÑÇÍÂèÒ§´éÇÂÍѵÃÒ¤ÇÒÁ¶Õ à·èÒ ¡Ñº
â´ÂÊÁÁØµÔ ÇèÒ Ç§¨Ã¡Òêѡ µÑÇÍÂèÒ§ÁÕ ¡ÒÃà¢éÒ ¨Ñ§ËÇÐẺ
ÊÁºÙóì (perfect synchronization) ¨Ò¡¹Ñ ¹ ÅӴѺ¢éÍÁÙÅ·Õ ä´é
{sk }
¨Ð¶Ù¡Êè§ä»ÂѧÍÕ¤ÇÍäÅà«ÍÃì à¾× Í
»ÃѺ ÃÙ»ÃèÒ§¢Í§ÊÑÒ³ãËé à» ¹ 仵ÒÁ·ÒÃìà¡çµ ·Õ µéͧ¡Òà ¹Ñ ¹¤×Í ÍÕ¤ÇÍäÅà«ÍÃì ¨Ð¾ÂÒÂÒÁ·ÓãËé ÅӴѺ ¢éÍÁÙÅàÍÒµì¾Øµ
{yk }
·Õ ä´éÁÕ¤èÒã¡Åéà¤Õ§¡ÑºÅӴѺ¢éÍÁÙÅ·Õ µéͧ¡ÒÃ
µÃǨËÒÇÕà·ÍÃìºÔ¡ç¨Ð·Ó¡ÒöʹÃËÑÊ¢éÍÁÙÅ
{yk }
{dk }
ãËéÁÒ¡·Õ ÊØ´ ËÅѧ¨Ò¡¹Ñ ¹ ǧ¨Ã
à¾× ÍËÒÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
{ak }
·Õ à» ¹ä»ä´éÁÒ¡·Õ ÊØ´
㹡Ò÷´Åͧ ¤èÒ SNR ¨Ð¹ÔÂÒÁµÒÁÊÁ¡Òà (3.27) ¹Ñ ¹¤×Í
à SNR = 10 log10
â´Â·Õ
Vp = g(∞) = 1
pulse) ³ àÇÅÒ
t
=
∞
Vp2 σ2
! (dB)
¤×Í ¢¹Ò´¢Í§ÊÑÒ³¾ÑÅÊì à»ÅÕ Â¹Ê¶Ò¹ÐàÍ¡à·È
áÅÐ
σ2
=
N0 /(2T )
¤×Í ¡ÓÅѧ¢Í§ÊÑҳú¡Ç¹
(isolated transition
n(t)
㹡Ò÷´Åͧ¹Õ ·ÒÃìà¡çµáÅÐÍÕ¤ÇÍäÅà«ÍÃì¨Ð¶Ù¡Í͡Ẻ·Õ ND = 2.5 áÅÐ SNR = 22 dB ³ ÃдѺ¤ÇÒÁÃعáç¢Í§ÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡ µèÒ§æ àªè¹
2, d2effmin , σw
áÅÐ
SNReff
σj
µÒÁ·Õ ¡Ó˹´ÁÒãËé Êèǹ¾ÒÃÒÁÔàµÍÃì
à» ¹µé¹ ¨Ð¶Ù¡¤Ó¹Ç³ËÒâ´Âãªé¢éÍÁÙÅà¾Õ§ 1 à«¡àµÍÃì (4096
ºÔµ) ÊÓËÃѺ áµèÅÐ SNR áÅÐ ND ã¹·Õ ¹Õ ¨Ð¾Ô¨ÒóÒ੾ÒзÒÃìà¡çµ ·Õ ¶Ù¡ Í͡Ẻâ´Âà§× ͹䢺ѧ¤Ñº ẺâÁ¹Ô¡ (´ÙÃÒÂÅÐàÍÕ´ã¹ËÑÇ¢éÍ·Õ 3.2.1) áÅÐà¾× ÍãËé§èÒµèÍ¡ÒÃ͸ԺÒ ÊÑÅѡɳì GPRn ¨Ðãªé á·¹·ÒÃìà¡çµáººâÁ¹Ô¡·Õ Áըӹǹá·ç»à·èҡѺ
n á·ç» ¹Í¡¨Ò¡¹Õ
㹡Ò÷´Åͧ à¤Ã× Í§ËÁÒ − ¨Ð
ãªéá·¹¢éÍÁÙÅ −2, à¤Ã× Í§ËÁÒ + ¨Ðãªéá·¹¢éÍÁÙÅ +2, áÅÐÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Ø¡µÑǨÐÁդسÊÁºÑµÔÊÁÁҵáѹ ¹Ñ ¹¤×Í
5.5.1
εa (D)
εa (D) = −εa (D)
¡ÒÃÇÔà¤ÃÒÐËìÃÐÂзҧ·Õ ¹éÍÂÊØ´
¨Ò¡áºº¨ÓÅͧáÅÐà§× Í¹ä¢·Õ ¡Ó˹´ãËé ¨Ðä´éÇèÒ ·ÒÃìà¡çµáºº GPR3 ·Õ ¶Ù¡Í͡ẺÊÓËÃѺÃкº¡Òà ºÑ¹·Ö¡áººá¹ÇµÑ §·Õ ND = 2.5 áÅÐ
σj = 0%
¤×Í
H(D) = 1 + 1.3022D + 0.6623D2
¨Ò¡¹Ñ ¹
ãËé ãªé ·ÒÃìà¡çµ ¹Õ 㹡Ò÷ӡÒèÓÅͧÃкº (system simulation) à¾× Í ¤Ó¹Ç³ËÒÃÐÂзҧÂؤÅÔ´ ¡ÓÅѧ Êͧ
d2 {εa (D)},
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧÊͧ
d2eff {εa (D)},
áÅÐÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ
εa (D)
µèÒ§æ ·Õ à¡Ô´¢Ö ¹ ã¹Ãкº (·Õ ´éÒ¹¢ÒÍÍ¡¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ) ³ SNR = 22 dB «Ö §¨Ðä´éµÒÁ·Õ
5.5.
¼Å¡Ò÷´Åͧ
105
µÒÃÒ§·Õ 5.1: ÃÐÂзҧÂؤÅÔ´ ¡ÓÅѧ Êͧ áÅÐÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ
εa (D)
d2 {εa (D)},
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧ Êͧ
d2eff {εa (D)},
¢Í§Ãкº·Õ ãªé·ÒÃìà¡çµáºº GPR3 ·Õ ND = 2.5 áÅÐ SNR =
22 dB
¤ÇÒÁÂÒÇ (ºÔµ) ¢Í§
ÃÐÂзҧÂؤÅÔ´ (Euclidean distance)
d2
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å (E ective distance)
ÅӴѺ¢éͼԴ¾ÅÒ´
¨Ó¹Ç¹¤ÃÑ §
εa (D)
·Õ à¡Ô´¢Ö ¹
¢éͼԴ¾ÅÒ´ 1
12.5375
2
7.7579
+
3
8.2764
4
d2eff
ÅӴѺ¢éͼԴ¾ÅÒ´
¨Ó¹Ç¹¤ÃÑ §
εa (D)
·Õ à¡Ô´¢Ö ¹
0
17.0522
127
6.7243
+
2
10.0266
+
2
8.7950
+ +
1
11.2688
+ +
1
5
9.3135
+ +
0
7.0754
+ 0 +
6
9.8321
+ + +
0
9.5535
+ 0 +
1
7
10.3507
+ + +
0
10.6952
+ + 0 +
1
8
10.8692
+ + + +
0
7.8234
+ 0 + 0 +
4
Í× ¹æ
0
+
127
37
50
7
áÊ´§ã¹µÒÃÒ§·Õ 5.1 ¶éÒãªéËÅÑ¡¤ÇÒÁ¨ÃÔ§·Õ ÇèÒ ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ ÁÕ¤èÒÃÐÂзҧ·Õ ¹éÍÂÊØ´¨Ðà» ¹ÅӴѺ¢éͼԴ¾ÅÒ´ ÍÔ¹¾Øµ ·Õ à¡Ô´ ¢Ö ¹ ºèÍÂ·Õ ÊØ´ ´Ñ§¹Ñ ¹ ¨Ò¡µÒÃÒ§·Õ 5.1 ¶éÒ ãªé ÃÐÂзҧÂؤÅÔ´ ¡ÓÅѧ Êͧ ࡳ±ì 㹡ÒþԨÒÃ³Ò ¨Ðä´é ÇèÒ ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Õ à¡Ô´ ¢Ö ¹ ºèÍ ¤×Í
{
+ +} à¹× ͧ¨Ò¡ ÁÕ ¤èÒ
d2 {εa (D)}
ãªéÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧÊͧ ÍÔ¹¾Øµ·Õ à¡Ô´¢Ö ¹ºèÍ ¤×Í
+},
{
à» ¹
+ }, áÅÐ
à·èÒ ¡Ñº 7.7579, 8.2764, áÅÐ 8.7950 µÒÁÅӴѺ áµè ¶éÒ
d2eff {εa (D)}
{ +}, {
{
d2 {εa (D)}
à» ¹à¡³±ì㹡ÒþԨÒÃ³Ò ¨Ðä´éÇèÒ ÅӴѺ¢éͼԴ¾ÅÒ´
+ 0 +}, áÅÐ
{
+ 0 + 0 +} à¹× ͧ¨Ò¡ ÁÕ¤èÒ
d2eff {εa (D)}
à·èҡѺ 6.7243, 7.0754, áÅÐ 7.8234 µÒÁÅӴѺ à¾× Íà» ¹¡ÒõÃǨÊͺ´ÙÇèÒ à¡³±ìã´ (ÃÐÂзҧÂؤÅÔ´ ËÃ×ÍÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å) ÁÕ¤ÇÒÁ¹èÒàª× Ͷ×Í ÁÒ¡¡Çèҡѹ ÊÓËÃѺãªé㹡Ò÷ӹÒÂÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ à¡Ô´¢Ö ¹ºèÍÂã¹Ãкº ¨Ö§ä´é·Ó¡ÒèÓÅͧ ÃкºÍÕ¡ ¤ÃÑ § Ë¹Ö § â´Â¡ÒÃÊè§ ¢éÍÁÙÅ ËÅÒÂæ à«¡àµÍÃì à¢éÒ ä»ã¹Ãкº (µÒÁẺ¨ÓÅͧã¹ÃÙ» ·Õ 3.2) ·Õ
106
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
SNR = 22 dB ¨¹¡ÃÐ·Ñ § ¢éͼԴ¾ÅÒ´ (error) ÊÐÊÁ·Õ à¡Ô´ ·Ñ §ËÁ´ÁÕ à» ¹ ¨Ó¹Ç¹ 500 ºÔµ ¨Ò¡¹Ñ ¹ ·Ó¡ÒÃÇÔà¤ÃÒÐËìÇèÒ ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾ØµáµèÅÐẺà¡Ô´¢Ö ¹·Ñ §ËÁ´¡Õ ¤ÃÑ § áÅÐàÁ× ÍÊÔ ¹ÊØ´¡Ò÷´Åͧ ¨Ð¾ºÇèÒ ÃкºÁÕ BER =
5.1945 × 10−4 ,
ºÔµ·Õ ¼Ô´¾ÅÒ´Áըӹǹ 500 ºÔµ, áÅÐÅӴѺ¢éͼԴ¾ÅÒ´
ÍÔ¹¾Øµ ·Ñ §ËÁ´ÁÕ ¨Ó¹Ç¹ 180 µÑÇ (ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ áµèÅÐẺ ÍÒ¨¨Ðà¡Ô´ ¢Ö ¹ ä´é ËÅÒ¤ÃÑ § ËÃ×Í ËÅÒµÑÇ) ¨Ò¡µÒÃÒ§·Õ 5.1 ¨Ð¾ºÇèÒ ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Õ à¡Ô´ ¢Ö ¹ ºèÍ ¤×Í +}, áÅÐ
{
+ 0 + 0 +} µÒÁÅӴѺ «Ö § ÊÍ´¤Åéͧ¡Ñº ¡ÒÃãªé ࡳ±ì
·Ó¹Ò¼Šà¾ÃÒÐ©Ð¹Ñ ¹
d2eff {εa (D)}
{
+},
{
d2eff {εa (D)}
+ 0
㹡ÒÃ
ÊÒÁÒö¹ÓÁÒãªé㹡Ò÷ӹÒÂÇèÒ ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµã´·Õ
à¡Ô´¢Ö ¹ºèÍÂã¹Ãкºä´é¶Ù¡µéͧÁÒ¡¡ÇèÒ¡ÒÃãªé
d2 {εa (D)}
ÍÂèÒ§äáçµÒÁ ¢éÍÊÃØ»¹Õ ¨ÐÁÕ¤ÇÒÁ¹èÒàª× Ͷ×Í
ÁÒ¡¢Ö ¹ ¡çµèÍàÁ× Í ã¹ÃкºÁÕÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ â´´à´è¹à¾Õ§µÑÇà´ÕÂÇ ¹Ñ ¹¤×Í àÁ× ÍÃкº·Ó§Ò¹ ·Õ ÃдѺ SNR ¤è͹¢éÒ§ÊÙ§ (ËÃ×Í·Õ ÃдѺ BER
< 10−4 )
㹷ӹͧà´ÕÂǡѹ µÒÃÒ§·Õ 5.2 áÊ´§¡ÒÃÇÔà¤ÃÒÐËìà˵ءÒóì¢éͼԴ¾ÅÒ´ÊÓËÃѺÃкº·Õ ãªé·ÒÃìà¡çµ Ẻ PR2,
H(D) = 1 + 2D + D2 ,
·Õ ND = 2.5 áÅÐ SNR = 22 dB «Ö §ã¹¡Ã³Õ¹Õ ¢éÍÁÙÅËÅÒÂæ
à«¡àµÍÃì¨Ð¶Ù¡Êè§à¢éÒä»ã¹Ãкº ¨¹¡ÃÐ·Ñ §ä´é BER =
1.4251 × 10−3 ,
ºÔµ·Õ ¼Ô´¾ÅÒ´Áըӹǹ 502
ºÔµ, áÅÐÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Ñ §ËÁ´ÁÕ ¨Ó¹Ç¹ 179 µÑÇ ¨ÐàËç¹ ä´é ÇèÒ ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Õ à¡Ô´ ¢Ö ¹ ºèÍÂã¹Ãкº¹Õ ¤×Í
d2eff {εa (D)}
{
+} áÅÐ
{
+ 0 +} µÒÁÅӴѺ «Ö § ÊÍ´¤Åéͧ¡Ñº ¡ÒÃãªé ࡳ±ì
㹡Ò÷ӹÒ¼Šà¾ÃÒÐ©Ð¹Ñ ¹ ¨Ò¡¼Å¡Ò÷´ÅͧÊÒÁÒöÊÃØ»ä´éÇèÒ ÅӴѺ¢éͼԴ¾ÅÒ´
ÍÔ¹¾Øµ·Õ à¡Ô´¢Ö ¹ºèÍ ¤×Í ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ ÁÕ ÍÔ¹¾Øµ·Õ ÁÕ
d2 {εa (D)}
d2eff {εa (D)}
¹éÍÂ·Õ ÊØ´ (äÁèãªèÅӴѺ¢éͼԴ¾ÅÒ´
¹éÍÂ·Õ ÊØ´) ¹Í¡¨Ò¡¹Õ ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ ÊÍ´¤Åéͧ¡Ñº
äÁè¨Óà» ¹¨Ðµéͧ໠¹µÑÇà´ÕÂǡѹ¡Ñº ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ ÊÍ´¤Åéͧ¡Ñº
d2eff {εa (D)}
d2 {εa (D)}
ã¹ÊèǹµèÍä»¹Õ ¨Ð·Ó¡ÒÃÇÔà¤ÃÒÐËì¼Å¡Ãзº¢Í§ÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡µèÍ¡ÒÃà¡Ô´ ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµã¹Ãкº ãËé¾Ô¨ÒóÒÃкº¡Òúѹ·Ö¡áººá¹ÇµÑ §·Õ ND = 2.5 â´Âãªé·ÒÃìà¡çµ Ẻ GPR5 (¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§áµèÅÐá·ç» ¢Í§·ÒÃìà¡çµ ÊÓËÃѺ áµèÅÐ ã¹ÃÙ» ·Õ 3.7(b)) µÒÃÒ§·Õ 5.3 áÊ´§ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ àËÅèÒ¹Ñ ¹ ³ ¨Ø´ ·Õ ÃкºÁÕ BER = ¢Í§
σj /T
¹éÍ (0%
Ẻá¹ÇµÑ § ¤×Í
− 3%)
{2, −2}
10−4
εa (D)
σj /T
ÊÒÁÒö´Ù ä´é ¨Ò¡¢éÍÁÙÅ
áÅФÇÒÁ¶Õ 㹡ÒÃà¡Ô´ ÅӴѺ
¨Ò¡µÒÃÒ§·Õ 5.3 àÁ× Í ÃдѺ ¤ÇÒÁÃعáç
¨ÐàËç¹ä´éÇèÒ ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ â´´à´è¹ÊÓËÃѺÃкº¡Òúѹ·Ö¡
¹Í¡¨Ò¡¹Õ ¨Ó¹Ç¹¢Í§ÅӴѺ ¢éͼԴ¾ÅÒ´·Õ â´´à´è¹ ¨ÐÁÕ à¾Ô Á ¢Ö ¹ àÁ× Í ÃдѺ
5.5.
¼Å¡Ò÷´Åͧ
107
µÒÃÒ§·Õ 5.2: ÃÐÂзҧÂؤÅÔ´ ¡ÓÅѧ Êͧ áÅÐÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ¤ÇÒÁÂÒÇ (ºÔµ) ¢Í§
εa (D)
d2 {εa (D)},
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧ Êͧ
¢Í§Ãкº·Õ ãªé·ÒÃìà¡çµáºº PR2 ·Õ ND=2.5 áÅÐ SNR=22 dB
ÃÐÂзҧÂؤÅÔ´ (Euclidean distance)
2
d
d2eff {εa (D)},
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å (E ective distance)
ÅӴѺ¢éͼԴ¾ÅÒ´
¨Ó¹Ç¹¤ÃÑ §
d2eff
ÅӴѺ¢éͼԴ¾ÅÒ´
¨Ó¹Ç¹¤ÃÑ §
εa (D)
·Õ à¡Ô´¢Ö ¹
{εa (D)}
εa (D)
·Õ à¡Ô´¢Ö ¹
0
34.0001
127
12.7191
+
¢éͼԴ¾ÅÒ´
{εa (D)}
1
24
2
16
+
3
16
+
9
16.1802
+
9
4
16
+ +
0
19.3624
+ +
0
5
16
+ +
2
13.6519
+ 0 +
6
16
+ + +
1
16.6625
+ 0 +
2
7
16
+ + +
0
16.8916
+ + +
0
8
16
+ + + +
0
15.5885
+ 0 + 0 +
2
Í× ¹æ
¤ÇÒÁÃعáç¢Í§
40
σj /T
0 127
26
13
ÁÒ¡¢Ö ¹ ÊÓËÃѺ ¡ÒÃÇÔà¤ÃÒÐËì à˵ءÒÃ³ì ¢éͼԴ¾ÅÒ´¢Í§Ãкº¡Òúѹ·Ö¡ Ẻ
á¹Ç¹Í¹ ¨Ð¾ºÇèÒ ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ â´´à´è¹ ¤×Í
{2, −2, 2}
[19]
»ÃÐ⪹ì·Õ ä´é¨Ò¡¡ÒÃÈÖ¡ÉÒ¡ÒÃÇÔà¤ÃÒÐËìà˵ءÒóì¢éͼԴ¾ÅÒ´ ¤×Í àÁ× Í·ÃÒºÇèÒÃкº·Õ ãªéÁÕÅӴѺ ¢éͼԴ¾ÅÒ´ ÍÔ¹¾Øµ ·Õ â´´à´è¹ Ẻ ã´ ¹Ñ¡ÇԨѠÊÒÁÒö ·Õ ¨Ð à¾Ô Á »ÃÐÊÔ·¸ÔÀÒ¾ ÃÇÁ ¢Í§ Ãкº ä´é â´Â ¡Òà Í͡ẺÃËÑÊ RLL (run length limited) [9] ËÃ×Íǧ¨Ãà¢éÒÃËÑÊ¡è͹ (precoder) [47] à¾× Íãªéà¢éÒÃËÑÊ ¢éÍÁÙÅ¢èÒÇÊÒáè͹·Õ ¨Ð·Ó¡ÒÃà¢Õ¹ŧä»ã¹Ê× ÍºÑ¹·Ö¡ à¾× ÍËÅÕ¡àÅÕ Â§¡ÒÃà¡Ô´ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ â´´à´è¹àËÅèÒ¹Ñ ¹ ÃÐËÇèÒ§¡Ãкǹ¡ÒöʹÃËÑÊ¢éÍÁÙÅ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
5.5.2
¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§
ã¹ËÑÇ¢éÍ¹Õ ¨ÐáÊ´§ãËéàËç¹ÇèÒ
SNReff
SNReff
ààÅÐ BER
áÅÐ BER ÁÕ¤ÇÒÁÊÑÁ¾Ñ¹¸ì¡Ñ¹¤è͹¢éÒ§ÁÒ¡ â´Â੾ÒÐÍÂèÒ§ÂÔ §
àÁ× Í¤ÇÒÁÃعáç¢Í§ÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡
σj /T
ÁÕ¤èÒ¹éÍ áÅÐÃкº·Ó§Ò¹·Õ ÃдѺ
SNR ÊÙ§à¾Õ§¾Í(¹Ñ ¹¤×Í àÁ× Íã¹ÃкºÁÕà˵ءÒóì¢éͼԴ¾ÅÒ´·Õ â´´à´è¹à¾Õ§µÑÇà´ÕÂÇ)
108
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
µÒÃÒ§·Õ 5.3: ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ÃкºÁÕ BER =
εa (D)
¢Í§Ãкº·Õ ãªé·ÒÃìà¡çµáÅÐ
σj /T
ẺµèÒ§æ ³ ¨Ø´·Õ
10−4
ÅӴѺ
PR2
¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ
σj /T
+
GPR5
= 0%
σj /T
GPR5
= 0%
σj /T
GPR5
= 3%
σj /T
= 6%
σj /T
= 9%
4.90%
3.19%
3.36%
5.84%
41.53%
67.54%
83.25%
79.66%
35.21%
9.62%
+ +
5.79%
0.35%
1.98%
38.31%
21.53%
+ +
0.51%
0.58%
1.14%
6.66%
15.73%
+ + +
0.13%
0.23%
0.48%
2.66%
6.31%
+ 0 +
15.53%
8.75%
8.94%
3.03%
0.00%
+ 0 + 0 +
1.34%
0.73%
0.84%
0.22%
0.00%
Í× ¹æ
4.26%
2.72%
3.60%
8.06%
5.28%
+
ÃÙ»·Õ 5.5 à»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº·Õ ãªé·ÒÃìà¡çµáºº GPR5 ³ ÃдѺ ÃÙ»¢Í§ BER áÅÐ
SNReff
SNReff
σj /T
·Õ ND = 2.5 ¨ÐàËç¹ä´éÇèÒ »ÃÐÊÔ·¸ÔÀÒ¾ã¹ÃÙ»¢Í§ BER áÅÐ
¼ÅÅѾ¸ìà» ¹ä»µÒÁ·Õ ¤Ò´ËÇѧäÇé ¡ÅèÒǤ×Í àÁ× Í ¤èÒ
GPR5
σj /T
µèÒ§æ ã¹
SNReff
ÁÕ
ÁÕ¤èÒà¾Ô Á¢Ö ¹ ¤èÒ BER ¢Í§Ãкº¡ç¨Ðà¾Ô Á¢Ö ¹ áÅÐ
¢Í§Ãкº¡ç¨ÐÅ´¹éÍÂŧ ¹Í¡¨Ò¡¹Õ ÃÙ»·Õ 5.5(a) áÊ´§ãËéàËç¹ÇèÒ ³ ÃдѺ¤ÇÒÁÃعáç
¢Í§ÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡µ Ó ¤èÒ
SNReff
ÊÒÁÒö¹ÓÁÒãªé㹡Ò÷ӹÒ¤èÒ BER ä´é
¨Ò¡ÊÁ¡Òà (5.18) áÅÐ (5.20) ¨Ðä´éÇèÒ [19]
BER ≈ K2 Q â´Â·Õ
K2
¤×Í ¤èÒ¤§µÑÇ·Õ äÁè¢Ö ¹¡Ñº
2 σw
µ p ¶ 1 SNReff 2
µÑÇÍÂèÒ§àªè¹ àÁ× Í
σj /T = 0%
(5.22)
¤èÒ BER ·Õ ·Ó¹ÒÂä´é¨ÐáÊ´§
´éÇÂàÊé¹ Q(·) ã¹ÃÙ» «Ö §ÊÍ´¤Åéͧ¡Ñº¤èÒ BER ¨ÃÔ§·Õ ä´é¨Ò¡¡ÒèÓÅͧÃкº àÁ× Í ÃÙ»·Õ 5.6(a) áÊ´§¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§ BER áÅРẺ㴡çµÒÁ ¶éÒÃкºÁÕ¤èÒ àÁ× Í
σj /T
SNReff
ÁÕ ¤èÒ ¹éÍ ´Ñ§¹Ñ ¹ ¤èÒ
SNReff
K2 = 2.3
¨ÐàËç¹ä´éÇèÒ äÁèÇèÒÃкº¨Ðãªé·ÒÃìà¡çµ
à·èҡѹáÅéÇ Ãкº¡ç¨ÐÁÕ¤èÒ BER ·Õ ã¡Åéà¤Õ§¡Ñ¹ â´Â੾ÒÐÍÂèÒ§ÂÔ §
SNReff
¨Ö§ ÊÒÁÒö¹ÓÁÒãªé 㹡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§
·ÒÃìà¡çµ ẺµèÒ§æ ä´é àªè¹à´ÕÂǡѺ ¡ÒÃãªé ¤èÒ BER ÍÂèÒ§äáçµÒÁ ¤ÇõÃÐ˹ѡ äÇé ÇèÒ ÃкºµèÒ§æ ¨Ð
5.5.
¼Å¡Ò÷´Åͧ
109
−1
10
−2
BER
10
−3
10
jitter = 0% jitter = 3% jitter = 6% jitter = 9% Q(⋅) with jitter = 0%
−4
10
−5
10
14
15
16
17
(a)
18
19
20
21
22
23
22
23
Electronics SNR (dB)
19
18
17
SNReff (dB)
16
15
14
13
jitter = 0% jitter = 3% jitter = 6% jitter = 9%
12
11
10
9 14
15
16
17
(b)
18
19
20
21
Electronics SNR (dB)
ÃÙ»·Õ 5.5: »ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкºã¹ÃÙ» (a) BER áÅÐ (b)
SNReff
¢Í§·ÒÃìà¡çµáºº GPR5
ãªé »ÃÔÁÒ³ SNR ·Õ µèÒ§¡Ñ¹ 㹡Ò÷ÓãËé à¡Ô´ ¤èÒ BER áÅФèÒ
SNReff
·Õ à·èÒ ¡Ñ¹ ´Ñ§·Õ áÊ´§ã¹ÃÙ» ·Õ
5.6(b)
110
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
−1
10
−2
BER
10
−3
10
PR2 [1 2 1], jitter=0% GPR5, jitter=0% GPR5, jitter=3% GPR5, jitter=6%
−4
10
−5
10
8
10
12
14
(a)
16
18
20
Effective SNR (dB)
−1
10
−2
BER
10
−3
10
−4
10
PR2 [1 2 1], jitter=0% GPR5, jitter=0% GPR5, jitter=3% GPR5, jitter=6%
−5
10
14
15
16
17
(b)
ÃÙ»·Õ 5.6:
(a) ¡ÃÒ¿ BER áÅÐ
SNReff ,
18
19
20
21
22
23
Electronics SNR (dB)
(b) ¡ÃÒ¿ BER áÅÐ SNR ¢Í§Ãкº·Õ ãªé ·ÒÃìà¡çµ Ẻ
µèÒ§æ ³ ÃдѺ¤ÇÒÁÃعáç¢Í§ÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡·Õ à¢Õ¹ÇèÒ jitter µèÒ§æ ·Õ ND = 2.5
5.6.
ÊÃØ»·éÒº·
5.6
111
ÊÃØ»·éÒº·
㹡ÒÃÍ͡Ẻ·ÒÃìà¡çµ µÒÁà§× ͹䢺ѧ¤Ñº ẺâÁ¹Ô¡ (monic constraint) â´Âãªé Ẻ¨ÓÅͧã¹ÃÙ» ·Õ 3.2 ÅӴѺ ¢éͼԴ¾ÅÒ´
wk
ÊÒÁÒö·Õ ¨Ð¶Ù¡ ¾Ô¨ÒóÒä´é ÇèÒ à» ¹ ÊÑҳú¡Ç¹ã¹áºº¨ÓÅͧªèͧ
ÊÑÒ³ GPR ẺÊÁÁÙÅ µÒÁÃÙ»·Õ 5.2 ¶éÒÊÑҳú¡Ç¹
wk
à» ¹ÊÑҳú¡Ç¹à¡ÒÊìÊÕ¢ÒÇẺ
ºÇ¡áÅéÇ »ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº¨Ð¢Ö ¹ÍÂÙè¡Ñº ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ ÁÕ¤èÒÃÐÂзҧÂؤÅÔ´·Õ ¹éÍÂÊØ´
dmin
ÍÂèÒ§äáçµÒÁã¹·Ò§»¯ÔºÑµÔ
wk
ÁÑ¡¨ÐÁÕÅѡɳÐà» ¹ÊÑҳú¡Ç¹áººÊÕ â´Â੾ÒÐÍÂèÒ§ÂÔ §
·Õ ¤ÇÒÁ¨Ø ¢éÍÁÙÅ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ìÊÙ§æ ´Ñ§¹Ñ ¹ ã¹¡Ã³Õ ¹Õ »ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº¨Ð¢Ö ¹ ÍÂÙè ¡Ñº ÅӴѺ ¢éͼԴ¾ÅÒ´·Õ ÁÕ¤èÒÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å·Õ ¹éÍÂÊØ´
deffmin
¼Å¡Ò÷´ÅͧáÊ´§ãËéàËç¹ÇèÒ ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ à¡Ô´¢Ö ¹ºèÍ ¤×Í ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Õ ÁÕ
d2eff {εa (D)}
·Õ ¹éÍÂÊØ´ (äÁè ãªè ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Õ ÁÕ
d2 {εa (D)}
·Õ ¹éÍÂÊØ´) ¹Í¡¨Ò¡¹Õ
¨Ó¹Ç¹¢Í§ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ â´´à´è¹¨ÐÁÕà¾Ô Á¢Ö ¹ àÁ× ÍÃдѺ¤ÇÒÁÃعáç¢Í§ÊÑҳú¡Ç¹ ¨Ôµ àµÍÃì ¢Í§Ê× Í ºÑ¹·Ö¡ à´ÕÂÇ (àÁ× Í ÃкºÁÕ
σj /T
σj /T
ÁÒ¡¢Ö ¹ ã¹¡Ã³Õ ·Õ ÃкºÁÕ ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Õ â´´à´è¹ à¾Õ§µÑÇ
¹éÍ áÅзӧҹ·Õ ÃдѺ SNR ÊÙ§ à¾Õ§¾Í) ¹Ñ¡ÇԨѠÊÒÁÒöãªé»ÃÐ⪹ì
¨Ò¡¢éÍÁÙŢͧ¡ÒÃÇÔà¤ÃÒÐËìà˵ءÒóì¢éͼԴ¾ÅÒ´¹Õ ÁÒãªé㹡ÒÃÍ͡ẺÃËÑÊ RLL ËÃ×Íǧ¨Ãà¢éÒÃËÑÊ ¡è͹ à¾× Íãªéà¢éÒÃËÑÊ¢éÍÁÙÅ¢èÒÇÊÒáè͹·Õ ¨Ð·Ó¡ÒÃà¢Õ¹ŧä»ã¹Ê× ÍºÑ¹·Ö¡ à¾× ÍËÅÕ¡àÅÕ Â§¡ÒÃà¡Ô´ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Õ â´´à´è¹ àËÅèÒ¹Ñ ¹ ÃÐËÇèÒ§¡Ãкǹ¡ÒöʹÃËÑÊ ¢éÍÁÙÅ ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ «Ö §¨ÐªèÇ·ÓãËé»ÃÐÊÔ·¸ÔÀÒ¾ÃÇÁ¢Í§Ãкº´ÕÁÒ¡¢Ö ¹ ÊØ´·éÒÂä´éáÊ´§ãËéàËç¹ÇèÒ
SNReff
ãªé㹡ÒÃÇÑ´»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкºä´é àªè¹à´ÕÂǡѺ¡ÒÃãªé BER ¹Ñ ¹¤×Í Ãкº·Õ ÁÕ
ÊÒÁÒö¹ÓÁÒ
SNReff
¤èÒ BER µ Ó áÅж֧áÁéÇèÒ ÃкºáµèÅÐÃкº¨Ðãªé·ÒÃìà¡çµµèÒ§¡Ñ¹ áµè¶éÒÃкºàËÅèÒ¹Ñ ¹ÁÕ
ÊÙ§ ¡ç¨ÐÁÕ
SNReff
à·èÒ
¡Ñ¹ Ãкº¡ç¨ÐÁÕ BER ã¡Åéà¤Õ§¡Ñ¹´éÇÂ
5.7
à຺½ ¡ËÑ´·éÒº·
1. ¾Ô¨ÒóÒẺ¨ÓÅͧªèͧÊÑÒ³ PR2,
H(D) = 1 + 2D + D2 ,
ÍÔ¹¾ØµÁÕ¤ÇÒÁÂÒÇà·èҡѺ 25 ºÔµ â´Â·Õ ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ·Õ Ë¹Ö §
−1,
1, 1,
−1, −1,
1, 1,
−1,
1,
−1,
1, 1,
−1,
1,
−1,
1, 1,
¶éÒ ¡Ó˹´ãËé ÅӴѺ ¢éÍÁÙÅ
A1 (D) −1,
¤×Í
1, 1,
{1, −1, −1}
1, 1,
áÅÐÅӴѺ
112
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¢éÍÁÙÅÍÔ¹¾Øµ·Õ Êͧ 1, 1,
−1,
1, 1,
A2 (D)
−1, −1,
¤×Í
{1,
1, 1,
1,
−1}
−1,
1,
−1,
1,
−1,
−1,
1,
1, 1,
−1, −1,
1,
−1,
¨§ËÒÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ ¶Ù¡µéͧ·Ñ §ËÁ´·Õ ¾ºã¹
Ãкº¹Õ
2. ¡Ó˹´ãËé ÊÁÒªÔ¡ áµèÅеÑÇ ¢Í§ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ ÁÕ ¤èÒ ·Õ à» ¹ ä»ä´é ¤×Í ºÑ¹·Ö¡ Ẻá¹Ç¹Í¹
(longitudinal
recording)
{−1,
1} ã¹Ãкº¡ÒÃ
ÅӴѺ ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ ·Õ à¡Ô´ ¢Ö ¹ ºèÍÂ
{2, −2, 2}
ã¹ÃÐËÇèÒ§¡Ãкǹ¡ÒöʹÃËÑÊ ¢éÍÁÙÅ ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¤×Í
¨§¤Ó¹Ç³
ËÒà˵ءÒóì¢éͼԴ¾ÅÒ´¢Í§Ãкº¹Õ àÁ× ÍÃкºãªé·ÒÃìà¡çµ´Ñ§µèÍ仹Õ
2.1)
H(D) = 1 − D
2.2)
H(D) = 1 + D − D2 − D3
2.3)
H(D) = 1 + 2D − 2D3 − D4
2.4)
H(D) = 1 − 0.04D − 0.64D2
2.5)
H(D) = 1 + 0.22D − 0.65D2 − 0.36D3
2.6)
H(D) = 1 + 0.24D − 0.50D2 − 0.40D3 − 0.21D4
3. ¡Ó˹´ãËé ÊÁÒªÔ¡ áµèÅеÑÇ ¢Í§ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ ÁÕ ¤èÒ ·Õ à» ¹ ä»ä´é ¤×Í ºÑ¹·Ö¡áººá¹ÇµÑ §
(perpendicular
recording)
{−1,
1} ã¹Ãкº¡ÒÃ
ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ à¡Ô´¢Ö ¹ºèÍÂã¹
ÃÐËÇèÒ§¡Ãкǹ¡ÒöʹÃËÑÊ ¢éÍÁÙÅ ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
¤×Í
à˵ءÒóì¢éͼԴ¾ÅÒ´¢Í§Ãкº¹Õ àÁ× ÍÃкºãªé·ÒÃìà¡çµ´Ñ§µèÍ仹Õ
3.1)
H(D) = 1 + D
3.2)
H(D) = 1 + 3D + 3D2 + D3
3.3)
H(D) = 1 + 4D + 6D2 + 4D3 + D4
3.4)
H(D) = 1 + 1.30D + 0.66D2
3.5)
H(D) = 1 + 1.19D + 0.60D2 + 0.12D3
3.6)
H(D) = 1 + 1.21D + 0.62D2 + 0.16D3 + 0.01D4
{2, −2}
¨§¤Ó¹Ç³ËÒ
5.7.
à຺½ ¡ËÑ´·éÒº·
113
4. ¨§¾ÔÊÙ¨¹ì¤èÒÃÐÂзҧÂؤÅÔ´¡ÓÅѧÊͧ
d2 {εa (D)}
·Õ áÊ´§ã¹µÒÃÒ§·Õ 5.1
5. ¨§¾ÔÊÙ¨¹ì¤èÒÃÐÂзҧÂؤÅÔ´¡ÓÅѧÊͧ
d2 {εa (D)}
·Õ áÊ´§ã¹µÒÃÒ§·Õ 5.2
6. ¾Ô¨ÒóÒẺ¨ÓÅͧ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµã¹ÃÙ»·Õ 3.2 ÊÓËÃѺÃкº¡Òúѹ·Ö¡áººá¹Ç¹Í¹ ·Õ
ak ∈ {−1, 1}
ND = 2 áÅÐ SNR = 22 dB â´Â·Õ ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
{wk } = {−1.95,
1.66, 0.36,
·Õ à¡Ô´¢Ö ¹ºèÍÂÃкº ¤×Í
−0.63,
{2, −2, 2}
¶éÒÅӴѺ¢éͼԴ¾ÅÒ´
1.90, 0.10, 2.12} áÅÐÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ
¨§¤Ó¹Ç³ËÒÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧÊͧ
¢Í§ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ¹Õ àÁ× ÍÃкºãªé·ÒÃìà¡çµ
6.1)
H(D) = 1 − D
6.2)
H(D) = 1 − D2
6.3)
H(D) = 1 + D − D2 − D3
H(D)
εa (D)
d2eff {εa (D)}
´Ñ§µèÍ仹Õ
7. ¾Ô¨ÒóÒẺ¨ÓÅͧ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ ã¹ÃÙ» ·Õ 3.2 ÊÓËÃѺ Ãкº¡Òúѹ·Ö¡ Ẻá¹ÇµÑ § ·Õ ND = 2 áÅÐ SNR = 22 dB â´Â·Õ ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
{wk }
=
εa (D)
{−5.37, −4.65,
0.56, 5.60,
·Õ à¡Ô´ ¢Ö ¹ ºèÍÂÃкº ¤×Í
−2.40, −8.26,
{2, −2}
H(D) = 1 + D
7.2)
H(D) = 1 + 2D + D2
7.3)
H(D) = 1 + 3D + 3D2 + D3
¶éÒÅӴѺ¢éͼԴ¾ÅÒ´
4.85} áÅÐÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ
¨§ËÒÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧ Êͧ
¢Í§ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ¹Õ àÁ× ÍÃкºãªé·ÒÃìà¡çµ
7.1)
ak ∈ {−1, 1}
H(D)
´Ñ§µèÍ仹Õ
d2eff {εa (D)}
114
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
º··Õ 6
ǧ¨ÃµÃǨËÒ NPML
㹺·¹Õ ¨Ð͸ԺÒÂËÅÑ¡¡Ò÷ӧҹáÅлÃÐâª¹ì ¢Í§ ǧ¨ÃµÃǨËÒ NPML (noise predictive maxi mum likelihood) [51, 52] «Ö §à» ¹Ç§¨ÃµÃǨËÒ·Õ ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡¡ÇèÒǧ¨ÃµÃǨËÒ PRML (par tial response maximum likelihood) â´Â੾ÒÐÍÂèÒ§ÂÔ § àÁ× Í Ãкº·Ó§Ò¹·Õ ¤ÇÒÁ¨Ø ¢éÍÁÙÅ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ÊÙ§ ã¹·Ò§»¯ÔºÑµÔ ǧ¨ÃµÃǨËÒ NPML »ÃÐÂØ¡µì ÁҨҡǧ¨ÃµÃǨËÒ PRML â´Â¡ÒÃ¹Ó ¡Ãкǹ¡ÒÃ㹡Ò÷ӹÒÂÊÑҳú¡Ç¹ (noise prediction process) ὧà¢éÒ ä»ÍÂÙè ã¹áµèÅÐàÊé¹ ÊÒ¢Ò (branch) ¢Í§á¼¹ÀÒ¾à·ÃÅÅÔÊ (trellis diagram) ´Ñ§·Õ ¨Ð͸ԺÒµèÍä»ã¹º·¹Õ ¾ÃéÍÁ·Ñ §áÊ´§ ¼Å¡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾ÃÐËÇèҧǧ¨ÃµÃǨËÒ PRML áÅÐǧ¨ÃµÃǨËÒ NPML
6.1
º·¹Ó
à·¤¹Ô¤ PRML ¤×Í ¡ÒÃãªé §Ò¹ÃèÇÁ¡Ñ¹ ÃÐËÇèÒ§ÍÕ¤ÇÍäÅà«ÍÃì Ẻ PR (partial response) áÅÐǧ¨Ã µÃǨËÒÇÕà·ÍÃìºÔ (Viterbi detector) «Ö § à» ¹ ·Õ ¹ÔÂÁãªé §Ò¹ÁÒ¡ã¹Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì µÒÁ·Õ ͸ԺÒÂ㹺··Õ 4 ãËé ¾Ô¨ÒóÒẺ¨ÓÅͧªèͧÊÑÒ³·Õ äÁè µèÍà¹× ͧ·Ò§àÇÅÒ áººÊÁÁÙÅ ã¹â´àÁ¹
D
µÒÁÃÙ» ·Õ 6.1 àÁ× Í
A(D)
N (D) ¤×Í ÊÑҳú¡Ç¹à¡ÒÊìÊÕ¢ÒÇẺºÇ¡ ¤×Í ·ÒÃìà¡çµ (target),
¤×Í ¢éÍÁÙÅ ºÔµ ÍÔ¹¾Øµ,
(AWGN),
C(D)
¤×Í ªèͧÊÑÒ³,
F (D) ¤×Í ÍÕ¤ÇÍäÅà«ÍÃìẺ
PR,
H(D)
Y (D) ¤×Í ¢éÍÁÙÅ·Õ ¨Ð¶Ù¡Êè§à¢éÒä»·Ó¡ÒöʹÃËÑÊ¢éÍÁÙÅ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ, 115
116
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
N(D)
F(D)
channel
H ( D) Y(D) C ( D)
A(D) C(D)
Aˆ ( D )
Viterbi detector
ÃÙ»·Õ 6.1: Ẻ¨ÓÅͧªèͧÊÑÒ³·Õ äÁèµèÍà¹× ͧ·Ò§àÇÅÒẺÊÁÁÙÅ
áÅÐ
Â(D)
¤×Í ¤èÒ»ÃÐÁÒ³¢Í§¢éÍÁÙźԵÍÔ¹¾Øµ
A(D)
¨Ò¡ÃÙ» ·Õ 6.1 ¢éÍÁÙÅ àÍÒµì¾Øµ ¢Í§ÍÕ¤ÇÍäÅà«ÍÃì
Y (D)
ÊÒÁÒöà¢Õ¹໠¹ ÊÁ¡Ò÷ҧ¤³ÔµÈÒʵÃì
ä´éµÒÁÊÁ¡Òà (4.6) ¹Ñ ¹¤×Í
H(D) Y (D) = A(D)H(D) + N (D) | {z } C(D) | {z } wanted signal
(6.1)
W (D)
â´Â·Õ
W (D)
¤×Í ÊÑҳú¡Ç¹·Õ ¨Ðà¢éÒä»ã¹Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ (µÒÁ·Õ ͸ԺÒÂã¹ËÑÇ¢éÍ·Õ 4.2.2)
â´Â·Ñ Çä» Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¨Ð·Ó§Ò¹ä´éÍÂèÒ§ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡·Õ ÊØ´ ¡çµèÍàÁ× Í
W (D)
ÁÕÅѡɳÐ
à» ¹ ÊÑҳú¡Ç¹à¡ÒÊì ÊÕ ¢ÒÇẺºÇ¡ ÍÂèÒ§äáçµÒÁã¹·Ò§»¯ÔºÑµÔ áÅéÇ (â´Â੾ÒÐÍÂèÒ§ÂÔ § àÁ× Í Ãкº·Ó§Ò¹·Õ ¤ÇÒÁ¨Ø ¢éÍÁÙÅ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ìÊÙ§æ) ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ (tap) ¹éÍ ãËéÁռŵͺʹͧàËÁ×͹¡ÑºªèͧÊÑÒ³ 1
¨ÐÁÕ ÅѡɳÐà» ¹ ÊÑҳú¡Ç¹áººÊÕ
C(D)
H(D)
·Õ ÁÕ ¨Ó¹Ç¹á·ç»
·Óä´éÂÒ¡ÁÒ¡ ´Ñ§¹Ñ ¹â´Â·Ñ Çä»
W (D)
(colored noise) [25] «Ö § ¨ÐÊè§ ¼Å·ÓãËé »ÃÐÊÔ·¸ÔÀÒ¾¡ÒÃ
·Ó§Ò¹¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ Ŵŧ à¾ÃÒÐ©Ð¹Ñ ¹ ¶éÒ µéͧ¡ÒÃãËé ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ÊÒÁÒö·Õ ¨Ð ·Ó§Ò¹ä´éÍÂèÒ§ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ÊÙ§ÊØ´àËÁ×͹à´ÔÁ ¹Ñ¡ÇԨѨеéͧËÒÇÔ¸Õ¡ÒÃã´ÇÔ¸Õ¡ÒÃË¹Ö §ã¹¡Ò÷ÓãËéͧ¤ì »ÃСͺ¢Í§ÊÑҳú¡Ç¹ã¹¢éÍÁÙÅ
Y (D)
(¹Ñ ¹¤×Í
W (D))
ÁÕÅѡɳÐà» ¹ÊÑҳú¡Ç¹à¡ÒÊì
ÊÕ¢ÒÇẺºÇ¡ ¡è͹·Õ ¨ÐÊ觼ÅÅѾ¸ì·Õ ä´éà¢éÒä»·Ó¡ÒöʹÃËÑÊ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ´Ñ§¹Ñ ¹ÍÒ¨¨Ð¡ÅèÒÇä´éÇèÒ à·¤¹Ô¤ NPML [51, 52] ¤×Í ¡ÒÃãªé§Ò¹ÃèÇÁ¡Ñ¹ÃÐËÇèÒ§ÍÕ¤ÇÍäÅà«ÍÃìẺ PR áÅÐǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ·Õ ÁÕ¡Ãкǹ¡ÒÃ㹡Ò÷ӹÒÂÊÑҳú¡Ç¹ ËÃ×Í ¡Ãкǹ¡ÒÃ㹡Òà ·ÓãËé ÊÑҳú¡Ç¹à» ¹ ÊÕ ¢ÒÇ (noise whitening process) ὧÍÂÙè ¢éÒ§ã¹ÍÑÅ¡ÍÃÔ·ÖÁ ÇÕà·ÍÃìºÔ ´Ñ§ 1
¢éÍÁÙÅá«Á໠ŢͧÊÑҳú¡Ç¹
W (D)
áµèÅÐá«Á໠ŨÐÁÕÊËÊÑÁ¾Ñ¹¸ì (correlation) ¡Ñ¹
6.2.
¡Ãкǹ¡ÒÃ㹡Ò÷ӹÒÂÊÑҳú¡Ç¹
117
wk target
ak
H(D)
rk
yk
âk Viterbi algorithm
predictor
P(D) ÃÙ»·Õ 6.2: Ẻ¨ÓÅͧªèͧÊÑÒ³¾ÃéÍÁǧ¨ÃµÃǨËÒ NPML
áÊ´§ã¹ÃÙ»·Õ 6.2 à¾ÃÒÐ©Ð¹Ñ ¹ àÁ× ÍÃкº·Ó§Ò¹·Õ ¤ÇÒÁ¨Ø¢éÍÁÙŢͧÎÒÃì´´ÔÊ¡ìä´Ã¿ìÊÙ§æ ǧ¨ÃµÃǨËÒ NPML ¨Ö§ ¤ÇÃ·Õ ¨Ð¶Ù¡ ¹ÓÁÒãªé §Ò¹ÁÒ¡¡ÇèÒ ¡ÒÃãªé ǧ¨ÃµÃǨËÒ PRML à¾× Í·Õ ¨Ðä´é »ÃÐÊÔ·¸ÔÀÒ¾ÃÇÁ ¢Í§Ãкº·Õ ´Õ¡ÇèÒ
6.2
¡Ãкǹ¡ÒÃ㹡Ò÷ӹÒÂÊÑҳú¡Ç¹
ǧ¨Ã¡Ãͧ·Ó¹Ò (predictor lter) ·Õ ãªé§Ò¹·Ñ Çä»ã¹Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³´Ô¨Ô·ÑŨÐÁÕÅѡɳР·Ñ §·Õ à» ¹ Ẻ¼ÅµÍºÊ¹Í§ÍÔÁ¾ÑÅÊì ¨Ó¡Ñ´ (FIR: nite impulse response) áÅÐẺ¼ÅµÍºÊ¹Í§ ÍÔÁ¾ÑÅÊìäÁè¨Ó¡Ñ´ (IIR: in nite impulse response) ¹Í¡¨Ò¡¹Õ Õ ¤Ø³ÊÁºÑµÔ·Ñ Ç仢ͧǧ¨Ã¡Ãͧ·Ó¹Ò ¤×Í ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒ (prediction error) ¨Ð¤èÍÂæ Ŵŧ àÁ× Í¨Ó¹Ç¹á·ç»¢Í§Ç§¨Ã¡Ãͧ·Ó¹Ò à¾Ô Á¢Ö ¹ ã¹Ë¹Ñ§Ê×ÍàÅèÁ¹Õ ¨Ð¾Ô¨ÒóÒ੾ÒÐǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ FIR à·èÒ¹Ñ ¹ ¾Ô¨ÒóÒẺ¨ÓÅͧ¡Ãкǹ¡ÒÃ㹡Ò÷ÓãËéÊÑҳú¡Ç¹à» ¹ÊÕ¢ÒÇ ã¹ÃÙ»·Õ 6.3 àÁ× Í ÊÑҳú¡Ç¹áººÊÕ, áÅÐ
P (D)
ŵk
¤×Í ¤èÒ»ÃÐÁÒ³¢Í§
wk , ek = wk − ŵk
¤×Í ¿ §¡ìªÑ¹¶èÒÂâ͹¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂã¹â´àÁ¹
P (D) =
N X k=1
D
wk
¤×Í
¤×Í ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒÂ,
«Ö §ÁÕÃÙ»ÊÁ¡Òà ¤×Í
pk Dk = p1 D + p2 D2 + . . . + pN DN
(6.2)
118
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
wk
ek predictor
wˆ k
P(D)
ÃÙ»·Õ 6.3: Ẻ¨ÓÅͧ¡Ãкǹ¡ÒÃ㹡Ò÷ÓãËéÊÑҳú¡Ç¹à» ¹ÊÕ¢ÒÇ
â´Â·Õ
pk
¤×Í ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô µÑÇ·Õ
¡Ãͧ·Ó¹Ò ´Ñ§¹Ñ ¹
ŵk
k
¢Í§Ç§¨Ã¡Ãͧ·Ó¹Ò áÅÐ
N
¤×Í ¨Ó¹Ç¹á·ç»·Ñ §ËÁ´¢Í§Ç§¨Ã
ÊÒÁÒöà¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§ÊÁ¡Ò÷ҧ¤³ÔµÈÒʵÃìä´é ¤×Í
ŵk =
N X
pi wk−i
(6.3)
i=1 ÊÁ¡Òà (6.3) ¨ÐÃÙé¨Ñ¡¡Ñ¹ã¹ª× ͧ͢ ǧ¨Ã¡Ãͧ·Ó¹ÒÂË¹Ö §¢Ñ ¹áººàªÔ§àÊé¹ (linear one step predic tor) àªè¹à´ÕÂǡѹ ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒÂ
ek
ÊÒÁÒö¨Ñ´ ãËé ÍÂÙè ã¹ÃÙ» ¢Í§ÊÁ¡Ò÷ҧ¤³ÔµÈÒʵÃì ä´é
´Ñ§¹Õ
ek = wk − ŵk = wk −
N X
pi wk−i
(6.4)
i=1 ËÃ×Íà¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§â´àÁ¹
D
ä´é ¤×Í
E(D) = [1 − P (D)]W (D) â´Â·Õ ¾¨¹ì
[1 − P (D)]
(6.5)
¨ÐàÃÕ¡¡Ñ¹ ·Ñ Çä»ÇèÒ Ç§¨Ã¡Ãͧ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒ (prediction error
lter)
6.3
¡ÒÃËÒ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂ
¨Ø´»ÃÐʧ¤ì 㹡ÒÃÍ͡Ẻǧ¨Ã¡Ãͧ·Ó¹Ò ·Ó¹ÒÂ) ¤×Í ¡Ò÷ÓãËé¢éͼԴ¾ÅÒ´¡Ò÷ӹÒÂ
P (D) ek
(¹Ñ ¹¤×Í ¡ÒÃËÒ¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§Ç§¨Ã¡Ãͧ
ÁÕ¤èÒ¹éÍÂ·Õ ÊØ´ ËÃ×ÍÍÕ¡¹ÑÂË¹Ö §¡ç¤×Í ¡Ò÷ÓãËé
ek
ÁÕ
6.3.
¡ÒÃËÒ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂ
119
ÅѡɳÐà» ¹ ÊÑҳú¡Ç¹ÊÕ ¢ÒÇãËé ÁÒ¡·Õ ÊØ´ ·Ñ §¹Õ à» ¹ à¾ÃÒÐÇèÒ ¢éÍÁÙÅ
ek
¶×Í ÇèÒ à» ¹ ͧ¤ì»ÃСͺ
¢Í§ÊÑҳú¡Ç¹·Õ ËŧàËÅ×Í ÍÂÙè ã¹¢éÍÁÙÅ ·Õ ¨ÐÊè§ à¢éÒ ä»·Ó¡ÒöʹÃËÑÊ ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ´Ñ§¹Ñ ¹ ÇÔ¸Õ¡ÒÃËÒ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂ·Õ ´Õ·Õ ÊØ´ ¡ç¤×Í ¡Ò÷ÓãËé¤èÒ¢éͼԴ¾ÅÒ´¡ÓÅѧÊͧ à©ÅÕ Â
(MSE: mean squared error) [51, 52]
£ ¤ £ ¤ E e2k = E (wk − ŵk )2 ÁÕ ¤èÒ ¹éÍÂ·Õ ÊØ´ â´Â·Õ
E[·]
(6.6)
¤×Í µÑÇ´Óà¹Ô¹¡ÒäèÒ ¤Ò´ËÁÒ «Ö § ÊÒÁÒö·Óä´é â´Â¡ÒÃËÒ͹ؾѹ¸ì ¢Í§
ÊÁ¡Òà (6.6) à·Õº¡Ñº¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂ
pi
áµèÅеÑÇ áÅéÇãËé¼ÅÅѾ¸ì·Õ ä´éÁÕ¤èÒà·èÒ
¡Ñº¤èÒÈÙ¹Âì ¨Ò¡¹Ñ ¹ ·Ó¡ÒÃá¡éÃкºÊÁ¡ÒÃàªÔ§àÊ鹡ç¨Ðä´é¤ÓµÍºÍÍ¡ÁÒ ËÃ×Í ÍÒ¨¨ÐÍÒÈÑ ËÅÑ¡¡Òà àªÔ§µÑ §©Ò¡
(orthogonality principle) ·Õ ÇèÒ
E [(wk − ŵk )wm ] = 0 ÊÓËÃѺ
m = 1, 2, . . . , N
(6.7)
â´Â¡ÒÃá¡éÊÁ¡Òà (6.7) ¨Ðä´éÇèÒ
E[wk wm ] −
N X
pi E[wk−i wm ] = 0
i=1
E[wk wm ] = Rww (k − m) =
N X i=1 N X
pi E[wk−i wm ] pi Rww (k − i − m)
(6.8)
i=1 àÁ× Í
Rww (i)
¤×Í ¤èÒÍѵÊËÊÑÁ¾Ñ¹¸ì (auto correlation) ÅӴѺ·Õ
â´Â
Rww (i) = E[wk+i wk ] = E
"S−1 X
i
¢Í§ÊÑҳú¡Ç¹
wk
«Ö §¹ÔÂÒÁ
# wk+i wk
(6.9)
k=0 â´Â·Õ
S
¤×Í ¤ÇÒÁÂÒÇËÃ×ͨӹǹºÔµ¢Í§ÅӴѺ¢éÍÁÙÅ
{wk }
¶éÒá·¹¤èÒ
k−m=j
ã¹ÊÁ¡Òà (6.8)
¨Ðä´é¼ÅÅѾ¸ìà» ç¹ ÊÁ¡ÒùÍÃìÁÍÅ (normal equation) ¹Ñ ¹¤×Í
Rww (j) =
N X i=1
pi Rww (j − i)
(6.10)
120
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ÊÓËÃѺ
j = 1, 2, . . . , N
Rww (1)
Rww (2) . . . Rww (N ) | {z
= }
«Ö §ÊÒÁÒö¨Ñ´ãËéÍÂÙèã¹ÃÙ»¢Í§àÁ·ÃÔ¡«ìä´é ¤×Í
|
Rww (0)
Rww (1)
Rww (1)
Rww (0)
. . .
. . .
Rww (N − 1)
···
r
···
Rww (N − 1)
p1
Rww (N − 2) p2 .. . . . . . . . Rww (1) Rww (0) pN {z } | {z ···
(6.11)
}
p
R
ËÃ×Í
r = Rp à¹× ͧ¨Ò¡
R
(6.12)
à» ¹àÁ·ÃÔ¡«ì¨ÑµØÃÑÊ (square matrix) ´Ñ§¹Ñ ¹ ¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂ
p
ÊÒÁÒöËÒä´é¨Ò¡¡ÒÃá¡éÊÁ¡Òà (6.12) ¹Ñ ¹¤×Í
p = R−1 r áÅФèÒ ¢éͼԴ¾ÅÒ´¡ÓÅѧ Êͧà©ÅÕ Â ·Õ ¹éÍÂÊØ´
(6.13)
(MMSE: minimum mean squared error) ¢Í§Ç§¨Ã
¡Ãͧ·Ó¹Ò¨ÐÁÕ¤èÒà·èҡѺ [52]
N X £ 2¤ pi Rww (i) E ek = Rww (0) −
(6.14)
i=1
6.4
ËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒ NPML
à¾× ÍãËé§èÒµèÍ¡ÒÃ͸ԺÒÂËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒ NPML ãËé¾Ô¨ÒóÒẺ¨ÓÅͧªèͧÊÑÒ³ PR4 µÒÁÃÙ»·Õ 6.4 â´Â·Õ ¢éÍÁÙÅ·Õ ´éÒ¹¢Òà¢éҢͧǧ¨ÃµÃǨËÒ NPML ÊÒÁÒöà¢Õ¹ãËéÍÂÙèã¹ÃÙ»ÊÁ¡Òà ·Ò§¤³ÔµÈÒʵÃìä´é ¤×Í
yk = rk + wk = ak − ak−2 + wk àÁ× Í
rk = ak − ak−2
¤×Í ¢éÍÁÙÅàÍÒµì¾ØµªèͧÊÑÒ³ áÅÐ
wk
(6.15)
¤×Í ÊÑҳú¡Ç¹áººÊÕ
6.4.
ËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒ NPML
121
wk ak
PR4 target
1− D
yk
rk
2
NPML detector
aˆk
ÃÙ»·Õ 6.4: Ẻ¨ÓÅͧªèͧÊÑÒ³ PR4
ËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒ NPML ¨ÐµèÒ§¨Ò¡ËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ã¹ àÃ× Í§¢Í§¡ÒäӹdzàÁµÃÔ¡ÊÒ¢Ò (branch metric) ¡ÅèÒǤ×Í àÁµÃÔ¡ÊҢҢͧǧ¨ÃµÃǨËÒ NPML ¨Ð ÁÕ¾¨¹ì·Õ à» ¹¤èÒ·Ó¹Ò¢ͧÊÑҳú¡Ç¹
ŵk
à¢éÒÁÒÃèÇÁ´éÇ ¹Ñ ¹¤×Í
λk (u, q) = |yk − r̂k (u, q) − ŵk |2 àÁ× Í
r̂k (u, q)
¤×Í ¢éÍÁÙÅàÍÒµì¾ØµªèͧÊÑÒ³·Õ äÁèÁÕÊÑҳú¡Ç¹ (noiseless channel output) ·Õ
ÊÍ´¤Åéͧ¡Ñº¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð¨Ò¡Ê¶Ò¹Ð ³ àÇÅÒ·Õ
k
(6.16)
u
ä»Ê¶Ò¹Ð
q
¶éÒ¡Ó˹´ãËé
·Õ ÊÍ´¤Åéͧ¡ÑºàÊé¹·Ò§¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð¨Ò¡Ê¶Ò¹Ð
u
ak (q)
ä»Ê¶Ò¹Ð
¤×Í ¢éÍÁÙźԵÍÔ¹¾Øµ
q
´Ñ§¹Ñ ¹ ÊÓËÃѺªèͧ
ÊÑÒ³ PR4 ¨Ðä´éÇèÒ
r̂k (u, q) = ak (q) − ak−2 (q) ¹Í¡¨Ò¡¹Õ ¤èÒ ·Ó¹Ò¢ͧÊÑҳú¡Ç¹
ŵk
¤×Í
ŵk =
(6.17)
ÊÒÁÒöà¢Õ¹ãËé ÍÂÙè ã¹ÃÙ» ÊÁ¡Ò÷ҧ¤³ÔµÈÒʵÃì ä´é
N X
pi wk−i
(6.18)
i=1 á·¹¤èÒ
wk = yk − ak + ak−2
¨Ò¡ÊÁ¡Òà (6.15) ŧã¹ÊÁ¡Òà (6.18) ¨Ðä´é
ŵk =
N X
pi (yk−i − ak−i + ak−i−2 )
(6.19)
i=1 á·¹¤èÒ
r̂k (u, q)
¨Ò¡ÊÁ¡Òà (6.17) áÅÐ
ŵk
¨Ò¡ÊÁ¡Òà (6.19) ŧã¹ÊÁ¡Òà (6.16) ¨Ðä´éà» ¹
¯2 ¯ N ¯ ¯ X ¯ ¯ pi (yk−i − âk−i (q) + âk−i−2 (q))¯ λk (u, q) = ¯yk − ak (q) + ak−2 (q) − ¯ ¯ i=1
(6.20)
122
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
â´Â·Õ
âk (q)
¤×Í ¤èÒ»ÃÐÁÒ³¢Í§¢éÍÁÙźԵÍÔ¹¾Øµ ³ àÇÅÒ·Õ
(survivor path) ·Õ ÁҶ֧ʶҹÐ
k
·Õ ÊÍ´¤Åéͧ¡ÑºàÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè
q
àÁµÃÔ¡ÊÒ¢Òã¹ÊÁ¡Òà (6.20) äÁèàËÁÒÐÊÓËÃѺ¡ÒùÓÁÒãªé¡Ñº§Ò¹»ÃÐÂØ¡µì (application) ·Õ µéͧ ¡ÒäÇÒÁàÃçÇ ã¹¡ÒûÃÐÁÇżÅÊÙ§ àªè¹ ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì à¹× ͧ¨Ò¡ ¡ÒäӹdzàÁµÃÔ¡ ÊÒ¢ÒÁÕ ¿ §¡ìªÑ¹ ¡Òäٳ (ÊÓËÃѺ¡Ò÷ӹÒÂÊÑҳú¡Ç¹) à¾Ô Á¢Ö ¹ÁÒ á·¹·Õ ¨ÐÁÕ੾Òп §¡ìªÑ¹¡Òúǡ ¡ÒÃà»ÃÕº à·Õº ¡ÒÃàÅ×Í¡ (ACS: add compare select) àËÁ×͹¡Ñº·Õ ãªéã¹ÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔẺ¸ÃÃÁ´Ò ´Ñ§¹Ñ ¹ à¾× ÍãËéǧ¨ÃµÃǨËÒ NPML ÊÒÁÒö¹ÓÁÒãªé¡Ñº§Ò¹»ÃÐÂØ¡µì·Õ µéͧ¡ÒäÇÒÁàÃçÇ㹡ÒûÃÐÁÇżÅÊÙ§ ä´é ÊÁ¡Òà (6.20) ¨Ðµéͧ¶Ù¡¨Ñ´ÃÙ»ãËÁèãËéà» ¹
¯ ¯2 N +2 K ¯ ¯ X X ¯ ¯ λk (u, q) = ¯zk − ak−i (q)gi − ak (q)¯ âk−i (q)gi + ¯ ¯ i=1
i=K+1
àÁ× Í
K
(6.21)
¤×Í ¾ÒÃÒàµÍÃì·Õ ãªé㹡ÒûÃйջÃйÍÁÃÐËÇèÒ§¤ÇÒÁ«Ñº«é͹ (complexity) áÅлÃÐÊÔ·¸ÔÀÒ¾
¢Í§Ç§¨ÃµÃǨËÒ NPML ¡ÅèÒǤ×Í ¶éÒ
K
ÁÕ¤èÒÁÒ¡ ¤ÇÒÁ«Ñº«é͹¡ç¨ÐÁÒ¡ áµè»ÃÐÊÔ·¸ÔÀÒ¾·Õ ä´é¡ç¨Ð´Õ
(áÅÐã¹·Ò§µÃ§¡Ñ¹¢éÒÁ),
zk = yk −
N X
yk−i pi
(6.22)
i=1 ¤×Í ¢éÍÁÙÅàÍÒµì¾Øµ¢Í§Ç§¨Ã¡Ãͧ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒ ¤×Í ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§ ·ÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Å
[1 − P (D)]
´Ñ§áÊ´§ã¹ÃÙ»·Õ 6.5, áÅÐ
(e ective target) ã¹â´àÁ¹
D
gi
«Ö §¹ÔÂÒÁâ´Â
Heff (D) = 1 − g1 D − g2 D2 − . . . − gN +2 DN +2 = (1 − D2 )[1 − P (D)]
(6.23)
â´ÂÊÃØ»áÅéÇ ¡ÒÃÊÃéҧǧ¨ÃµÃǨËÒ NPML ã¹·Ò§»¯ÔºÑµÔ ·Óä´é´Ñ§µèÍ仹Õ
1) ¤Ó¹Ç³ËÒǧ¨Ã¡Ãͧ·Ó¹ÒÂ
2) ¤Ó¹Ç³ËÒÅӴѺ ¢éÍÁÙÅ
{yk }
{zk }
·Õ ÊÍ´¤Åéͧ¡Ñº·ÒÃìà¡çµ
P (D)]
µÒÁÃÙ»·Õ 6.5
P (D)
â´ÂãªéÊÁ¡Òà (6.13)
¨Ò¡ÊÁ¡Òà (6.22) â´Â¹ÓÅӴѺ ¢éÍÁÙÅ àÍÒµì¾Øµ ¢Í§ÍÕ¤ÇÍäÅà«ÍÃì
H(D)
·Õ µéͧ¡Òà ÁÒ¼èҹǧ¨Ã¡Ãͧ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒÂ
[1 −
6.4.
ËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒ NPML
123
yk
âk
zk Viterbi algorithm predictor
P(D) G(D)
ÃÙ»·Õ
6.5:
â¤Ã§ÊÃéÒ§ ¢Í§ ǧ¨Ã µÃǨËÒ
NPML
·Õ ãªé ¡Ñº §Ò¹ »ÃÐÂØ¡µì ·Õ µéͧ¡Òà ¤ÇÒÁ àÃçÇ ã¹ ¡ÒÃ
»ÃÐÁÇżÅÊÙ§
3) ¹ÓÅӴѺ ¢éÍÁÙÅ
{zk }
·Õ ä´é ä»·Ó¡ÒöʹÃËÑÊ ¢éÍÁÙÅ ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ Ẻ¸ÃÃÁ´Ò áµè
á¼¹ÀÒ¾à·ÃÅÅÔÊ ·Õ ãªé 㹡ÒäӹdzµÒÁÍÑÅ¡ÍÃÔ·ÖÁ ÇÕà·ÍÃìºÔ (µÒÁ·Õ ͸ԺÒÂã¹ËÑÇ¢éÍ ·Õ 4.3.3) ¨ÐµéͧÊÃéÒ§¨Ò¡·ÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Å
Heff (D)
«Ö §ËÒä´é¨Ò¡
Heff (D) = H(D)[1 − P (D)] àÁ× Í
H(D)
(6.24)
¤×Í ·ÒÃìà¡çµ·Õ ÊÍ´¤Åéͧ¡Ñº ÍÕ¤ÇÍäÅà«ÍÃì ·Õ ãªé ã¹Ãкº ¨ÐàËç¹ä´éÇèÒ ¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§
·ÒÃìà¡çµ »ÃÐÊÔ·¸Ô¼Å¨Ðà» ¹ àÅ¢¨Ó¹Ç¹¨ÃÔ§ ¡ÅèÒǤ×Í ¶Ö§áÁéÇèÒ ¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§ ¨Ó¹Ç¹àµçÁ áµè ¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§
P (D)
H(D)
à» ¹ àÅ¢ ¨Ó¹Ç¹¨ÃÔ§ ´Ñ§¹Ñ ¹ ¼ÅÅѾ¸ì ·Õ ä´é
¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô à» ¹ àÅ¢¨Ó¹Ç¹¨ÃÔ§ à¾ÃÒÐ©Ð¹Ñ ¹ ÍÒ¨¨Ð¡ÅèÒÇä´é ÇèÒ ·ÒÃìà¡çµ »ÃÐÊÔ·¸Ô¼Å
¨Ðà» ¹ àÅ¢
Heff (D) Heff (D)
¨ÐÁÕ ¤×Í
·ÒÃìà¡çµáºº GPR áººË¹Ö §¡çä´é
µÑÇÍÂèÒ§·Õ 6.1
¨Ò¡¡ÒÃÍ͡Ẻ·ÒÃìà¡çµáÅÐÍÕ¤ÇÍäÅà«ÍÃì µÒÁẺ¨ÓÅͧã¹ÃÙ»·Õ 3.2 ÊÓËÃѺÃкº
¡Òúѹ·Ö¡ Ẻá¹Ç¹Í¹ ·Õ ND = 2 áÅÐ SNR = 15 dB â´Â¡Ó˹´ãËé ·ÒÃìà¡çµ ·Õ µéͧ¡Òà ¤×Í
H(D) = 1 − D −0.66}
»ÃÒ¡®ÇèÒ
ÅӴѺ¢éͼԴ¾ÅÒ´
{wk }
·Õ ä´é ¤×Í
{0.86, −0.26, −0.13, −0.14, 0.35,
¨§¤Ó¹Ç³ËÒ
¡) ǧ¨Ã¡Ãͧ·Ó¹ÒÂ
P (D)
Ẻ 2 á·ç» áÅзÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Å
Heff (D)
124
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¢) ǧ¨Ã¡Ãͧ·Ó¹ÒÂ
ÇÔ¸Õ·Ó
P (D)
¨Ò¡ÅӴѺ¢éͼԴ¾ÅÒ´
Ẻ 4 á·ç» áÅзÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Å
{wk }
Heff (D)
·Õ ¡Ó˹´ãËé ¤èÒÍѵÊËÊÑÁ¾Ñ¹¸ì¢Í§
wk
ÊÒÁÒöËÒä´é¨Ò¡ÊÁ¡ÒÃ
(6.9) ´Ñ§¹Õ
Rww (i) = {0.2336, −0.0903, −0.0071, −0.0419, 0.2363, −0.5676} ÊÓËÃѺ
i = 0, 1, 2, . . . , 5
µÒÁÅӴѺ
¡) ǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 2 á·ç» ÊÒÁÒöËÒä´é¨Ò¡¡ÒÃá¡éÊÁ¡Òà (6.11) ¹Ñ ¹¤×Í
−0.0903
=
−0.0071
0.2336
−0.0903
−0.0903
0.2336
p1
p2
¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§Ç§¨Ã¡Ãͧ·Ó¹Ò ÁÕ¤èÒà·èҡѺ
p1 p2
=
−1 0.2336
−0.0903
−0.0903
0.2336 −0.0071 5.0323 1.9454 −0.0903 = 1.9454 5.0323 −0.0071 −0.4684 = −0.2116
−0.0903
´Ñ§¹Ñ ¹ ǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 2 á·ç» ¤×Í
P (D) = −0.4684D − 0.2116D2 áÅзÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Å·Õ ÊÍ´¤Åéͧ¡ÑºÇ§¨Ã¡Ãͧ·Ó¹ÒÂ¹Õ ¤×Í
£ ¤ Heff (D) = (1 − D) 1 − (−0.4684D − 0.2116D2 ) = 1 − 0.5316D − 0.2568D2 − 0.2116D3 â´Â¨Ó¹Ç¹Ê¶Ò¹Ðã¹á¼¹ÀÒ¾à·ÃÅÅÔÊ·Õ ãªéã¹Ç§¨ÃµÃǨËÒ NPML ¨ÐÁÕ·Ñ §ËÁ´
23 = 8
ʶҹÐ
6.4.
ËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒ NPML
125
¢) 㹷ӹͧà´ÕÂǡѹ ǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 4 á·ç» ¡çÊÒÁÒöËÒä´é¨Ò¡¡ÒÃá¡éÊÁ¡Òà (6.11) ¹Ñ ¹¤×Í
−0.0903
0.2336
−0.0903 −0.0071 −0.0419
p1
−0.0071 −0.0903 0.2336 −0.0903 −0.0071 p2 = −0.0419 −0.0071 −0.0903 0.2336 −0.0903 p3 0.2363 −0.0419 −0.0071 −0.0903 0.2336 p4 ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§Ç§¨Ã¡Ãͧ·Ó¹Ò ÁÕ¤èÒà·èҡѺ
p1
−1 0.2336
−0.0903 −0.0071 −0.0419
p2 −0.0903 0.2336 −0.0903 −0.0071 = p3 −0.0071 −0.0903 0.2336 −0.0903 p4 −0.0419 −0.0071 −0.0903 0.2336 5.9512 3.2140 2.2035 2.0163 −0.0903 3.2140 7.0038 3.6575 2.2035 −0.0071 = 2.2035 3.6575 7.0038 3.2140 −0.0419 2.0163 2.2035 3.2140 5.9512 0.2363 −0.1762 0.0274 = 0.2412 1.0739
−0.0903
−0.0071 −0.0419 0.2363
´Ñ§¹Ñ ¹ ǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 4 á·ç» ¤×Í
P (D) = −0.1762D + 0.0274D2 + 0.2412D3 + 1.0739D4 áÅзÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Å·Õ ÊÍ´¤Åéͧ¡ÑºÇ§¨Ã¡Ãͧ·Ó¹ÒÂ¹Õ ¤×Í
£ ¤ Heff (D) = (1 − D) 1 − (−0.1762D + 0.0274D2 + 0.2412D3 + 1.0739D4 ) = 1 − 0.8238D − 0.2036D2 − 0.2138D3 − 0.8327D4 + 1.0739D5
126
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
wk
target
ak
H(D)
yk
rk
NPML
zk
Viterbi
âk
G(D)
P(D) PRML
âk
Viterbi H(D)
ÃÙ»·Õ 6.6: Ẻ¨ÓÅͧªèͧÊÑҳẺÊÁÁÙÅ ¾ÃéÍÁ·Ñ §Ç§¨ÃµÃǨËÒ NPML áÅÐ PRML
â´Â¨Ó¹Ç¹Ê¶Ò¹Ðã¹á¼¹ÀÒ¾à·ÃÅÅÔÊ·Õ ãªéã¹Ç§¨ÃµÃǨËÒ NPML ¨ÐÁÕ·Ñ §ËÁ´
25 = 32
ʶҹÐ
Êѧࡵ¨Ð¾ºÇèÒ Ç§¨ÃµÃǨËÒ NPML ¨ÐÁÕ ¤ÇÒÁ«Ñº«é͹ÁÒ¡¡ÇèÒ Ç§¨ÃµÃǨËÒ PRML ·Ñ §¹Õ à» ¹ à¾ÃÒÐÇèÒ ·ÒÃìà¡çµ »ÃÐÊÔ·¸Ô¼Å
Heff (D)
¨ÐÁÕ ¨Ó¹Ç¹á·ç» ÁÒ¡¢Ö ¹ ¡ÇèÒ ·ÒÃìà¡çµ »¡µÔ
H(D)
«Ö § à» ¹ ¼Å
ÁÒ¨Ò¡¨Ó¹Ç¹á·ç» ¢Í§Ç§¨Ã¡Ãͧ·Ó¹Ò ËÃ×Í ÍÒ¨¨Ð¡ÅèÒÇä´é ÇèÒ ¨Ó¹Ç¹Ê¶Ò¹Ðã¹á¼¹ÀÒ¾à·ÃÅÅÔÊ ·Õ ãªéèã¹Ç§¨ÃµÃǨËÒ NPML Áըӹǹà·èҡѺ
¨Ó¹Ç¹Ê¶Ò¹Ð =
àÁ× Í »¡µÔ
|A|
|A|ν+N
á·¹¨Ó¹Ç¹¢éÍÁÙÅ ºÔµ ÍÔ¹¾Øµ ·Õ à» ¹ ä»ä´é ·Ñ §ËÁ´,
H(D),
áÅÐ
N
ν
¤×Í ¨Ó¹Ç¹Ë¹èǤÇÒÁ¨Ó¢Í§·ÒÃìà¡çµ
¤×Í ¨Ó¹Ç¹á·ç» ¢Í§Ç§¨Ã¡Ãͧ·Ó¹Ò ÍÂèÒ§äáçµÒÁ ¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨Ã
µÃǨËÒ NPML ÊÒÁÒö·ÓãËéŴŧä´éµÒÁ·Õ àʹÍã¹ [53]
µÑÇÍÂèÒ§·Õ 6.2
Ẻ¨ÓÅͧ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ áÅÐÍÕ¤ÇÍäÅà«ÍÃì ã¹ÃÙ» ·Õ 3.2 ÊÓËÃѺ Ãкº¡ÒÃ
ºÑ¹·Ö¡áººá¹Ç¹Í¹ ·Õ ND = 2 áÅÐ SNR = 15 dB ÊÒÁÒöŴÃÙ»ä´éà» ¹ Ẻ¨ÓÅͧªèͧÊÑÒ³ ẺÊÁÁÙÅ µÒÁÃÙ» ·Õ 6.6 àÁ× Í ·ÒÃìà¡çµ ·Õ µéͧ¡Òà ¤×Í
H(D) = 1 − D
¶éÒ ¡Ó˹´ãËé ÅӴѺ ¢éÍÁÙÅ
6.4.
ËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒ NPML
ÍÔ¹¾Øµ
{ak } = {−1, 1, 1, 1}
¨§¶Í´ÃËÑÊ¢éÍÁÙÅ
yk
127
áÅÐÊÑҳú¡Ç¹
{wk } = {0.46, −1.20, 1.02, −0.59, −0.98}
´éÇÂ
¡) ǧ¨ÃµÃǨËÒ PRML
¢) ǧ¨ÃµÃǨËÒ NPML àÁ× Íãªéǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 1 á·ç» ¤×Í
ÇÔ¸Õ·Ó
¨Ò¡ÃÙ»·Õ 6.6 ¢éÍÁÙÅàÍÒµì¾ØµªèͧÊÑÒ³
rk
P (D) = −0.753D
ËÒä´é¨Ò¡
rk = ak ∗ hk = {−1, 2, 0, 0, −1} ´Ñ§¹Ñ ¹ ¢éÍÁÙÅ·Õ ´éÒ¹¢Òà¢éҢͧǧ¨ÃµÃǨËÒ PRML áÅÐ NPML ¤×Í
yk = rk + wk = {−0.54, 0.80, 1.02, −0.59, −1.98} ¡)
ÊÓËÃѺ Ãкº PRML ǧ¨Ã µÃǨËÒ ÇÕà·ÍÃìºÔ ¨Ð ·Ó ¡Òà ¶Í´ÃËÑÊ ¢éÍÁÙÅ
à·ÃÅÅÔÊ ·Õ ÊÃéÒ§¨Ò¡·ÒÃìà¡çµ ¢éÍÁÙÅ
{yk }
H(D) = 1 − D
{yk }
â´Â ãªé á¼¹ÀÒ¾
µÒÁ·Õ áÊ´§ã¹ÃÙ» ·Õ 6.7(a) «Ö § ¢Ñ ¹µÍ¹¡ÒöʹÃËÑÊ
ÊÒÁÒöÊÃØ»ä´é µÒÁÃÙ»·Õ 6.8 â´Â·Õ µÑÇàÅ¢·Õ áÊ´§ÍÂÙ躹¨Ø´µèÍ (node) áµèÅШش ¤×Í ¤èÒ
àÁµÃÔ¡àÊé¹·Ò§·Õ ÁÒ¶Ö§ ³ ¨Ø´µèÍ¹Ñ ¹ áÅеÑÇàÅ¢·Õ áÊ´§ÍÂÙ躹àÊé¹ÊÒ¢ÒáµèÅÐàÊé¹ ¤×Í ¤èÒàÁµÃÔ¡ÊÒ¢Ò ¢Í§áµèÅÐàÊé¹ ÊÒ¢Ò·Õ ´Õ ·Õ ÊØ´ ·Õ ÁÒ¶Ö§ ·Õ ¨Ø´µèÍ ¹Ñ ¹æ à¾ÃÒÐ©Ð¹Ñ ¹ ¨Ò¡ÃÙ» ·Õ 6.8 ¤èÒ àÁµÃÔ¡ àÊé¹·Ò§·Õ ¹éÍ ·Õ ÊØ´ ¤×Í ¤èÒ 2.24 ´Ñ§¹Ñ ¹Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¨Ð¶Í´ÃËÑÊ¢éÍÁÙÅâ´Â¡ÒÃÁͧÂé͹¡ÅѺ仵ÒÁàÊé¹·Ò§ ·Õ Âѧ ÁÕ ªÕÇÔµ ÍÂÙè (survivor path) ·Õ ÁÒ¶Ö§ ³ ¨Ø´µèÍ ·Õ ÁÕ ¤èÒ àÁµÃÔ¡ àÊé¹·Ò§à·èÒ ¡Ñº 2.24 «Ö § ¨Ð¾ºÇèÒ ¤èÒ »ÃÐÁÒ³¢Í§ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
{âk }
·Õ ÊÍ´¤Åéͧ¡ÑºàÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè¹Õ ¤×Í
{âk } = {â0 , â1 , â2 , â3 } = {−1, −1, 1, 1} «Ö §ÁÕ¤èÒäÁèµÃ§¡ÑºÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
{ak } = {−1, 1, 1, 1}
·Õ Êè§ÁÒ¨Ò¡µé¹·Ò§ ´Ñ§¹Ñ ¹ ¡ÒöʹÃËÑÊ
´éÇÂǧ¨ÃµÃǨËÒ PRML ã¹¡Ã³Õ¹Õ ÁÕ¢éͼԴ¾ÅÒ´à¡Ô´¢Ö ¹à» ¹¨Ó¹Ç¹ 1 ºÔµ
¢)
ÊÓËÃѺÃкº NPML ÅӴѺ¢éÍÁÙÅ
{yk }
¨Ð¶Ù¡Ê觼èÒ¹à¢éÒä»ã¹Ç§¨Ã¡Ãͧ㹡Ò÷ÓãËéÊÑÒ³
ú¡Ç¹à» ¹ÊÕ¢ÒÇ (noise whitening lter) â´Â¨Ðä´é¼ÅÅѾ¸ìÍÍ¡ÁÒà» ¹ ËÃ×ÍáÊ´§à» ¹ÅӴѺ¢éÍÁÙÅ
{zk }
Z(D) = Y (D)[1 − P (D)]
ã¹â´àÁ¹àÇÅÒ ä´é´Ñ§¹Õ
{zk } = {−0.5400, 0.3934, 1.6224, 0.1781, −2.4243, −1.4910}
128
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
0
-1
0
-1 -1
2
2 -2
1
-1.5061
1 -1 0
0.4939
(a)
-0.4939
-1 1
ak = -1
1.5061
-2
ak = 1
1 1
0
(b) ÃÙ»·Õ 6.7: á¼¹ÀÒ¾à·ÃÅÅÔʢͧ (a) ·ÒÃìà¡çµ =
1 − 0.247D − 0.753D2
-1
0
0.29
0.29
H(D)
=
1−D
áÅÐ (b) ·ÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Å
·Õ ãªé㹡ÒöʹÃËÑÊ¢éÍÁÙŢͧÃкº PRML áÅÐ NPML µÒÁÅӴѺ
0.64
0.93
1.04
1.97
0.35
2.32
0.96
1
0
0.29
0.29
Heff (D)
0.64
0.93
2.24 0
1.89
0.35
2.24
3.92
6.16
ÃÙ»·Õ 6.8: á¼¹ÀÒ¾ÊÃØ»¢Ñ ¹µÍ¹¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔÊÓËÃѺÃкº PRML
¨Ò¡¹Ñ ¹ ÅӴѺ ¢éÍÁÙÅ
{zk }
¨Ð¶Ù¡ ¶Í´ÃËÑÊ ¢éÍÁÙÅ ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ·Õ ·Ó§Ò¹â´Âãªé á¼¹ÀÒ¾
à·ÃÅÅÔÊ·Õ ÊÃéÒ§¨Ò¡·ÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Å
Heff (D) = H(D)[1 − P (D)] = 1 − 0.247D − 0.753D2
µÒÁ·Õ áÊ´§ã¹ÃÙ» ·Õ 6.6(b) â´Â·Õ ¢Ñ ¹µÍ¹¡ÒöʹÃËÑÊ ¢éÍÁÙÅ
{zk }
ÊÒÁÒöÊÃØ» ä´é µÒÁÃÙ» ·Õ 6.9
6.5.
-1 -1
¼Å¡Ò÷´Åͧ
0
0.29
129
0.29
0.15
0.45
2.63
3.08
0.03
3.11
0.24
1.88
0.14 0.84
1 -1
0
1.07 0
0.59
0.01 0.79
4.59
9.56
0
4.18
0.45 4.48
-1 1
0 1.07
0
1.86
0.01
11
0
0.29
0.29
0.15
0.45
6.19 1.04
0.01 0.03
0.03
3.94
8.52
4.49 0.10
0.06
0.18 5.88
0.24
5.93
0.26
2.22
8.16
ÃÙ»·Õ 6.9: á¼¹ÀÒ¾ÊÃØ»¢Ñ ¹µÍ¹¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔÊÓËÃѺÃкº NPML
à¹× ͧ¨Ò¡ ¤èÒàÁµÃÔ¡ àÊé¹·Ò§·Õ ¹éÍÂ·Õ ÊØ´ ¤×Í ¤èÒ 0.24 à¾ÃÒÐ©Ð¹Ñ ¹ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¨Ð¶Í´ÃËÑÊ ¢éÍÁÙÅâ´Â¡ÒÃÁͧÂé͹¡ÅѺ仵ÒÁàÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè·Õ ÁÒ¶Ö§ ³ ¨Ø´µèÍ·Õ ÁÕ¤èÒàÁµÃÔ¡àÊé¹·Ò§à·èҡѺ 0.24 «Ö §¨Ð¾ºÇèÒ ¤èÒ»ÃÐÁÒ³¢Í§ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
{âk }
·Õ ÊÍ´¤Åéͧ¡ÑºàÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè¹Õ ¤×Í
{âk } = {â0 , â1 , â2 , â3 } = {−1, 1, 1, 1} «Ö §ÁÕ¤èҵç¡ÑºÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
{ak } = {−1, 1, 1, 1}
·Õ Êè§ÁÒ¨Ò¡µé¹·Ò§ ´Ñ§¹Ñ ¹ ¡ÒöʹÃËÑÊ´éÇÂ
ǧ¨ÃµÃǨËÒ NPML ã¹µÑÇÍÂèÒ§¢éÍ¹Õ ¨Ö§äÁèÁÕ¢éͼԴ¾ÅÒ´à¡Ô´¢Ö ¹
6.5
¼Å¡Ò÷´Åͧ
ã¹Êèǹ¹Õ ¨Ð·Ó¡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ç§¨ÃµÃǨËÒ PRML áÅÐ NPML â´ÂãªéẺ¨ÓÅͧ ªèͧÊÑÒ³¢Í§Ãкº¡Òúѹ·Ö¡ Ẻá¹Ç¹Í¹ (longitudinal recording) µÒÁÃÙ» ·Õ 6.10 â´Â·Õ
130
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
n(t)
ak
1− D {±1} 2
bk
p(t)
g(t)
sk
s(t) LPF
yk
equalizer
detector
âk
t k = kT ÃÙ»·Õ 6.10: Ẻ¨ÓÅͧªèͧÊÑÒ³¢Í§Ãкº¡Òúѹ·Ö¡áÁèàËÅç¡
ÊÑÒ³ read back ÊÒÁÒöà¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§ÊÁ¡Òä³ÔµÈÒʵÃìä´é ¤×Í
p(t) =
S−1 X
bk g(t − kT ) + n(t)
(6.25)
k=0
bk = (ak − ak−1 )/2
àÁ× Í
ºÇ¡ËÃ×Íź áÅÐ ·Ñ §ËÁ´
bk = 0
S = 4096
(1.1), áÅÐ
n(t)
´éÒ¹à·èҡѺ
N0 /2
¤×Í ºÔµà»ÅÕ Â¹Ê¶Ò¹Ð (bk
= ±1
ÊÍ´¤Åéͧ¡Ñº¡ÒÃà»ÅÕ Â¹á»Å§Ê¶Ò¹Ð
ËÁÒ¶֧ äÁèÁÕ¡ÒÃà»ÅÕ Â¹á»Å§Ê¶Ò¹Ð),
ºÔµ ËÃ×Í 1 à«¡àµÍÃì (sector),
g(t)
ak ∈ ±1
¤×Í ºÔµÍÔ¹¾Øµ·Õ Áըӹǹ
¤×Í ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹Ð µÒÁÊÁ¡ÒÃ
¤×Í ÊÑҳú¡Ç¹à¡ÒÊìÊÕ¢ÒÇẺºÇ¡·Õ ÁÕ¤ÇÒÁ˹Òá¹è¹Ê໡µÃÑÁ¡ÓÅѧẺÊͧ
ÊÑÒ³ read back
p(t)
¨Ð¶Ù¡Ê觼èÒ¹ä»Âѧǧ¨Ã¡Ãͧ¼èÒ¹µ Ӻѵà·ÍÃìàÇÔÃìµÍѹ´Ñº·Õ 7 áÅж١·Ó
¡ÒêѡµÑÇÍÂèÒ§´éǤÇÒÁ¶Õ ¡ÒêѡµÑÇÍÂèÒ§à·èҡѺ
1/T
â´ÂÊÁÁصÔÇèÒ ¡Ãкǹ¡ÒÃ㹡ÒêѡµÑÇÍÂèÒ§ÁÕ
¡ÒÃà¢éҨѧËÇÐÃÐËÇèÒ§ÊÑÒ³ read back áÅÐǧ¨ÃªÑ¡µÑÇÍÂèҧẺÊÁºÙóì (perfect synchroniza tion) ¨Ò¡¹Ñ ¹ ÅӴѺ¢éÍÁÙÅàÍÒµì¾Øµ
{sk }¨Ð¶Ù¡» ͹ä»ÂѧÍÕ¤ÇÍäÅà«ÍÃì (equalizer) à¾× Í»ÃѺÃÙ»ÃèÒ§¢Í§
ÊÑÒ³ãËéà» ¹ä»µÒÁ·ÒÃìà¡çµ·Õ µéͧ¡Òà áÅéÇ¡çÊè§ÅӴѺ¢éÍÁÙÅàÍÒµì¾Øµ
{yk }
·Õ ä´é ä»·Ó¡ÒöʹÃËÑÊ
¢éÍÁÙÅ ´éÇÂǧ¨ÃµÃǨËÒ (detector) à¾× Í ËÒ¤èÒ »ÃÐÁÒ³¢Í§ÅӴѺ ¢éÍÁÙÅ ÍÔ¹¾Øµ
{ak }
·Õ à» ¹ ä»ä´é ÁÒ¡
·Õ ÊØ´ ã¹·Õ ¹Õ ¤èÒ SNR ·Õ ãªé¨Ð¹ÔÂÒÁâ´Â
µ SNR = 10 log10 àÁ× Í
Vp = 1
áÅÐ
σ2
=
Vp 2 σ2
¶ (dB)
¤×Í ¢¹Ò´ÊÙ§ÊØ´¢Í§ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹ÐàÍ¡à·È
N0 /(2T )
¤×Í ¡ÓÅѧ¢Í§ÊÑҳú¡Ç¹
n(t)
(6.26)
(isolated transition pulse)
¹Í¡¨Ò¡¹Õ áµèÅШش¢Í§ÍѵÃÒ¢éͼԴ¾ÅÒ´
6.5.
¼Å¡Ò÷´Åͧ
131
−1
10
−2
BER
10
−3
10
−4
10
PRML: 22 states NPML (2−tap predictor): 24 states NPML (4−tap predictor): 26 states
−5
10
10
11
12
13
14
15
16
SNR (dB)
ÃÙ»·Õ 6.11: »ÃÐÊÔ·¸ÔÀÒ¾¢Í§ÃкºµèÒ§æ ·Õ ND = 2
ºÔµ (BER) ¨Ð¶Ù¡ ¤Ó¹Ç³â´Âãªé ¢éÍÁÙÅ ËÅÒÂæ à«¡àµÍÃì ¨¹¡ÇèÒ ¨Ðä´é ¢éͼԴ¾ÅÒ´ºÔµ ÁÒ¡¡ÇèÒ ËÃ×Í à·èÒ ¡Ñº 1000 ºÔµ ÃÙ» ·Õ 6.11 à»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ç§¨ÃµÃǨËÒ PRML áÅÐ NPML ·Õ ND = 2 àÁ× Í ¡Ó˹´ãËé·Ø¡Ãкºãªé·ÒÃìà¡çµáºº PR4,
H(D) = 1 − D2
ã¹á¼¹ÀÒ¾à·ÃÅÅÔʢͧǧ¨ÃµÃǨËÒ PRML ¤×Í
22 = 4
·Õ ãªéã¹á¼¹ÀÒ¾à·ÃÅÅÔʢͧǧ¨ÃµÃǨËÒ NPML ¤×Í áºº 2 á·ç») áÅÐ
22+4 = 64
à¾ÃÒÐ©Ð¹Ñ ¹ ¨Ó¹Ç¹Ê¶Ò¹Ð·Ñ §ËÁ´·Õ ãªé ʶҹР㹢³Ð·Õ ¨Ó¹Ç¹Ê¶Ò¹Ð·Ñ §ËÁ´
22+2 = 16
ʶҹР(ÊÓËÃѺǧ¨Ã¡Ãͧ·Ó¹ÒÂ
ʶҹР(ÊÓËÃѺǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 4 á·ç») ´Ñ§¹Ñ ¹¨ÐàËç¹ä´éÇèÒ
ǧ¨ÃµÃǨËÒ NPML ÁÕ ¤ÇÒÁ«Ñº«é͹ÁÒ¡¡ÇèÒ Ç§¨ÃµÃǨËÒ PRML áµè ¨Ò¡¼Å¡Ò÷´ÅͧµÒÁÃÙ» ·Õ 6.11 ¾ºÇèÒ Ç§¨ÃµÃǨËÒ NPML ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡¡ÇèÒǧ¨ÃµÃǨËÒ PRML ÍÂèÒ§àËç¹ä´éªÑ´ ËÃ×Í ÍÒ¨¨Ð¡ÅèÒÇä´éè ÇèÒ ³ ÃдѺ BER =
10−4
ǧ¨ÃµÃǨËÒ NPML ÁÕ »ÃÐÊÔ·¸ÔÀÒ¾´Õ ¡ÇèÒ Ç§¨ÃµÃǨËÒ
PRML »ÃÐÁÒ³ 2 dB ¹Í¡¨Ò¡¹Õ ¨Ò¡¼Å¡Ò÷´ÅͧÂѧ¾ºÇèÒ Ç§¨ÃµÃǨËÒ NPML ·Õ ãªéǧ¨Ã¡Ãͧ ·Ó¹Ò 2 á·ç» ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ã¡Åéà¤Õ§¡ÑºÇ§¨Ã¡Ãͧ·Ó¹Ò 4 á·ç» ´Ñ§¹Ñ ¹ ÊÓËÃѺÃкº·Õ ¾Ô¨ÒóҹÕ
132
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
0.078 0.076
Predictor MMSE
0.074 0.072 0.07 0.068 0.066 0.064 0.062
1
2
3
4
5
6
7
8
9
10
Number of predictor taps
ÃÙ»·Õ 6.12: »ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂ·Õ ãªé¨Ó¹Ç¹á·ç»µèÒ§¡Ñ¹ ·Õ SNR = 17 dB
ǧ¨ÃµÃǨËÒ NPML ÊÒÁÒöãªéǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 2 á·ç» ¡çà¾Õ§¾ÍµèÍ¡ÒÃãªé§Ò¹áÅéÇ à¹× ͧ¨Ò¡ ãËé»ÃÐÊÔ·¸ÔÀÒ¾·Õ ã¡Åéà¤Õ§¡Ñº¡ÒÃãªéǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 4 á·ç» áµè¤ÇÒÁ«Ñº«é͹¨Ð¹éÍ¡ÇèÒÁÒ¡ ÃÙ»·Õ 6.12 à»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂ·Õ ãªé¨Ó¹Ç¹á·ç»µèÒ§¡Ñ¹ ·Õ SNR = 17 dB ¨ÐàËç¹ä´éÇèÒ Ç§¨Ã¡Ãͧ·Ó¹Ò 2 á·ç» ¡çÁÕ»ÃÐÊÔ·¸ÔÀÒ¾à¾Õ§¾ÍÊÓËÃѺ¡ÒÃãªé§Ò¹áÅéÇ à¹× ͧ¨Ò¡ ¶Ö§áÁéÇèÒ¨Ðà¾Ô Á¨Ó¹Ç¹á·ç»ÁÒ¡¢Ö ¹ »ÃÐÊÔ·¸ÔÀÒ¾·Õ ä´é¡çà¾Ô Á¢Ö ¹¹éÍÂÁÒ¡ «Ö §äÁè¤ØéÁ¤èҡѺ¤ÇÒÁ«Ñº«é͹·Õ ä´éÃѺ 㹷ӹͧà´ÕÂǡѹÃÙ»·Õ 6.13 à»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ç§¨ÃµÃǨËÒ PRML áÅÐ NPM L ·Õ ND = 2.5 â´Âãªé·ÒÃìà¡çµáºº PR4 àËÁ×͹à´ÔÁ áµè¤ÃÒÇ¹Õ ¨Ð¾ºÇèÒ Ç§¨ÃµÃǨËÒ NPML ãËé »ÃÐÊÔ·¸ÔÀÒ¾·Õ ´Õ¡ÇèÒǧ¨ÃµÃǨËÒ PRML ¤è͹¢éÒ§ÁÒ¡ àÁ× Íà·Õº¡Ñº¡Ò÷ӧҹ·Õ ND = 2 â´ÂÍÒ¨¨Ð ¡ÅèÒÇä´éèÇèÒ ³ ÃдѺ BER =
10−4
ǧ¨ÃµÃǨËÒ NPML ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒǧ¨ÃµÃǨËÒ PRML
»ÃÐÁÒ³ 3 dB ·Ñ §¹Õ à» ¹à¾ÃÒÐÇèÒ ·Õ ¤ÇÒÁ¨Ø¢éÍÁÙŢͧÎÒÃì´´ÔÊ¡ìä´Ã¿ìÊÙ§ (ËÃ×Í ND ÊÙ§) ͧ¤ì»ÃСͺ ¢Í§ÊÑҳú¡Ç¹·Õ ὧÍÂÙè ã¹¢éÍÁÙÅ ·Õ ¨Ð·Ó¡ÒöʹÃËÑÊ ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨ÐÁÕ ÅÑ¡É³Ð à» ç¹ ÊÑҳú¡Ç¹áººÊÕ ÁÒ¡¢Ö ¹ ¨Ö§ ·ÓãËé¡ÒÃãªé §Ò¹Ç§¨ÃµÃǨËÒ NPML ä´é »ÃÐÊÔ·¸ÔÀÒ¾·Õ ´Õ ¡ÇèÒ
6.5.
¼Å¡Ò÷´Åͧ
133
−1
10
−2
BER
10
−3
10
−4
10
PRML: 22 states NPML (2−tap predictor): 24 states −5
10
12
13
14
15
16
17
SNR (dB)
ÃÙ»·Õ 6.13: »ÃÐÊÔ·¸ÔÀÒ¾¢Í§ÃкºµèÒ§æ ·Õ ND = 2.5
¡ÒÃãªé§Ò¹Ç§¨ÃµÃǨËÒ PRML ÁÒ¡ à¾× ÍãËé ¡ÒÃà»ÃÕºà·Õº໠¹ ä»ÍÂèÒ§ÂصԸÃÃÁã¹àÃ× Í§¢Í§¤ÇÒÁ«Ñº«é͹¢Í§Ãкº ¨Ð·Ó¡Ò÷´Åͧ à»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº 3 Ãкº ·Õ ND = 2.5 ´Ñ§µèÍ仹Õ
1) ǧ¨ÃµÃǨËÒ NPML ·Õ ãªé·ÒÃìà¡çµáºº PR4
H(D) = 1 − D2
áÅÐǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 2
á·ç»
2) ǧ¨ÃµÃǨËÒ PRML ·Õ ãªé·ÒÃìà¡çµáºº PR 2
3) ǧ¨ÃµÃǨËÒ GPRML
H(D) = 1 + 2D − 2D3 − D4
·Õ ãªé ·ÒÃìà¡çµ Ẻ GPR «Ö § Í͡Ẻâ´Âà§× ͹䢺ѧ¤Ñº ẺâÁ¹Ô¡ (´Ù
ÃÒÂÅÐàÍÕ´ã¹ËÑÇ¢éÍ·Õ 3.2.1) â´Â·Õ ·ÒÃìà¡çµáºº GPR ¹Õ ¨Ð¶Ù¡Í͡ẺÊÓËÃѺáµèÅÐ SNR
µÒÁÃÙ» ·Õ 6.14 àÁ× Íá¼¹ÀÒ¾à·ÃÅÅÔÊ ·Õ ãªé 㹡ÒöʹÃËÑÊ ¢éÍÁÙÅ ´éÇÂÍÑÅ¡ÍÃÔ·ÖÁ ÇÕà·ÍÃìºÔ ¢Í§·Ø¡ Ãкº 2
ǧ¨ÃµÃǨËÒ GPRML ¤×Í Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ·Õ ãªé ·ÒÃìà¡çµ Ẻ GPR ã¹¢³Ð·Õ ǧ¨ÃµÃǨËÒ PRML ¤×Í Ç§¨Ã
µÃǨËÒÇÕà·ÍÃìºÔ·Õ ãªé·ÒÃìà¡çµáºº PR
134
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
−1
10
2+2
NPML: 2 = 16 states PRML: [1 2 0 −2 −1] GPRML: (5−tap GPR) −2
BER
10
−3
10
−4
10
−5
10
12
13
14
15
16
17
SNR (dB)
ÃÙ»·Õ 6.14: ¼Å¡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§ÃкºµèÒ§æ ·Õ ND = 2.5
¨ÐÁÕ ¨Ó¹Ç¹Ê¶Ò¹Ðà·èÒ ¡Ñ¹ ¤×Í 16 ʶҹРáÅÐÍÕ¤ÇÍäÅà«ÍÃì ·Õ ãªé ¢Í§áµèÅÐÃкº¨Ð¶Ù¡ Í͡ẺãËé àËÁÒÐÊÁ¡Ñº·ÒÃìà¡çµ
H(D)
·Õ ¡Ó˹´ ¨Ò¡ÃÙ»¨ÐàËç¹ä´éÇèÒ Ç§¨ÃµÃǨËÒ NPML ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ÁÒ¡
¡ÇèÒ Ç§¨ÃµÃǨËÒ PRML ·Õ ãªé ·ÒÃìà¡çµ Ẻ PR áµè ÁÕ »ÃÐÊÔ·¸Ô ã¡Åéà¤Õ§¡Ñº ǧ¨ÃµÃǨËÒ GPRML ·Ñ §¹Õ à» ¹ à¾ÃÒÐÇèÒ ·ÒÃìà¡çµ »ÃÐÊÔ·¸Ô¼Å
Heff (D) = H(D)[1 − P (D)]
·Õ ãªé 㹡ÒÃÊÃéҧἹÀÒ¾
à·ÃÅÅÔʢͧǧ¨ÃµÃǨËÒ NPML ÊÒÁÒö·Õ ¨Ð¶Ù¡¾Ô¨ÒóÒä´éÇèÒà» ¹·ÒÃìà¡çµáºº GPR áººË¹Ö §ä´é à¹× ͧ¨Ò¡ ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§·ÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Åà» ¹àÅ¢¨Ó¹Ç¹¨ÃÔ§
6.6
ÊÃØ»·éÒº·
àÁ× Í Ãкº·Ó§Ò¹·Õ ¤ÇÒÁ¨Ø ¢éÍÁÙÅ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ÊÙ§ ͧ¤ì»ÃСͺ¢Í§ÊÑҳú¡Ç¹·Õ ´éÒ¹¢Òà¢éÒ ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ¨ÐÁÕÅѡɳÐà» ¹ÊÑҳú¡Ç¹áººÊÕ (colored noise) ÁÒ¡¢Ö ¹ ã¹¡Ã³Õ ¹Õ ǧ¨ÃµÃǨËÒ PRML äÁè ÊÒÁÒö·Ó§Ò¹ä´é ÍÂèÒ§ÁÕ »ÃÐÊÔ·¸ÔÀÒ¾ ´Ñ§¹Ñ ¹Ç§¨ÃµÃǨËÒ NPML ¨Ö§
6.7.
à຺½ ¡ËÑ´·éÒº·
135
ä´é ¶Ù¡ ¹ÓÁÒãªé à¾× Í à¾Ô Á »ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº ·Ñ §¹Õ à¹× ͧÁÒ¨Ò¡ÇèÒǧ¨ÃµÃǨËÒ NPML ¨ÐÁÕ ¡ÒÃãªé ǧ¨Ã¡Ãͧ·Ó¹Ò (㹡ÒÃ·Õ ¨Ð·ÓãËéÊÑҳú¡Ç¹áººÊÕ¡ÅÒÂà» ¹ÊÑҳú¡Ç¹ÊÕ¢ÒÇ) ÃèÇÁ¡Ñº ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ㹡ÒöʹÃËÑÊ ¢éÍÁÙÅ ¹Í¡¨Ò¡¹Õ Âѧ ¾ºÇèÒ·ÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Å·Õ ãªé 㹡ÒÃÊÃéÒ§ á¼¹ÀÒ¾à·ÃÅÅÔÊ ¢Í§Ç§¨ÃµÃǨËÒ NPML ÊÒÁÒö·Õ ¨Ð¶Ù¡ ¾Ô¨ÒóÒä´é ÇèÒ à» ¹ ·ÒÃìà¡çµ Ẻ GPR ¨Ö§ à» ¹à˵ؼŢéÍË¹Ö §ÇèÒ·ÓäÁǧ¨ÃµÃǨËÒ NPML ¨Ö§ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒǧ¨ÃµÃǨËÒ PRML à¾ÃÒÐÇèÒ ·ÒÃìà¡çµáºº GPR ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒ·ÒÃìà¡çµáºº PR µÒÁ·Õ ͸ԺÒÂ㹺··Õ 3 ¶Ö§áÁéÇèÒǧ¨ÃµÃǨËÒ NPML ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒǧ¨ÃµÃǨËÒ PRML áµèǧ¨ÃµÃǨËÒ NPML ¨ÐÁÕ¤ÇÒÁ«Ñº«é͹ÁÒ¡¡ÇèÒ à¾ÃÒÐ©Ð¹Ñ ¹ 㹡ÒõѴÊÔ¹ã¨ÇèҨйÓǧ¨ÃµÃǨËÒ NPML ÁÒãªé§Ò¹ËÃ×Í äÁè ãËé¾Ô¨ÒóÒÇèÒ »ÃÐÊÔ·¸ÔÀÒ¾·Õ ¨Ðä´éÃѺà¾Ô Á¢Ö ¹¨Ð¤ØéÁ¤èҡѺ¤ÇÒÁ«Ñº«é͹·Õ µÒÁÁÒËÃ×ÍäÁè
6.7
à຺½ ¡ËÑ´·éÒº·
1. ¨§Í¸ÔºÒÂ·Õ ÁҢͧá¹Ç¤Ô´¢Í§Ç§¨ÃµÃǨËÒ NPML
2. ¨§Í¸ÔºÒ¤ÇÒÁᵡµèÒ§¢Í§Ç§¨ÃµÃǨËÒ PRML áÅÐǧ¨ÃµÃǨËÒ NPML
3. ¨Ò¡¡ÒÃÍ͡Ẻ·ÒÃìà¡çµáÅÐÍÕ¤ÇÍäÅà«ÍÃì µÒÁẺ¨ÓÅͧã¹ÃÙ»·Õ 3.2 ÊÓËÃѺÃкº¡Òúѹ·Ö¡ Ẻá¹ÇµÑ § (perpendicular recording) ·Õ ND = 2.5 áÅÐ SNR = 20 dB â´Â¡Ó˹´ ãËé ·ÒÃìà¡çµ ·Õ µéͧ¡Òà ¤×Í
H(D) = 1 + D
{1.56, 0.35, −0.66, −0.69, 0.81, 0.20} ·Ó¹ÒÂ
P (D),
·ÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Å
3.1) ǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 1 á·ç» 3.2) ǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 2 á·ç» 3.3) ǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 3 á·ç» 3.4) ǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 4 á·ç»
â´Â ·Õ
Heff (D),
¢éÍÁÙÅ ¢Í§Ç§¨ÃµÃǨËÒ NPML ·Õ ãªé
»ÃÒ¡®ÇèÒ ÅӴѺ ¢éͼԴ¾ÅÒ´
ak ∈ {−1, 1}
{wk }
·Õ ä´é ¤×Í
¨§ ¤Ó¹Ç³ ËÒ Ç§¨Ã ¡Ãͧ
áÅÐáÊ´§á¼¹ÀÒ¾à·ÃÅÅÔÊ·Õ ãªé㹡ÒöʹÃËÑÊ
136
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
4. ·ÓàËÁ×͹㹢éÍ 3 áµè ¡Ó˹´ãËé ·ÒÃìà¡çµ ·Õ µéͧ¡Òà ¤×Í ¢éͼԴ¾ÅÒ´
H(D) = 1 + 2D + D2
áÅÐÅӴѺ
{wk } = {−0.56, −1.65, −0.21, 0.49, −0.98, −0.09}
5. Ẻ¨ÓÅͧ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ áÅÐÍÕ¤ÇÍäÅà«ÍÃì ã¹ÃÙ» ·Õ 3.2 ÊÓËÃѺ Ãкº¡Òúѹ·Ö¡ Ẻ á¹ÇµÑ § ·Õ ND = 2.5 áÅÐ SNR = 20 dB ÊÒÁÒö·Õ ¨ÐÅ´ÃÙ»ä´éà» ¹áºº¨ÓÅͧªèͧÊÑÒ³ ẺÊÁÁÙÅ µÒÁÃÙ»·Õ 6.6 àÁ× Í ·ÒÃìà¡çµ·Õ µéͧ¡Òà ¤×Í ÍÔ¹¾Øµ
H(D) = 1+D ¶éÒ¡Ó˹´ãËé ÅӴѺ¢éÍÁÙÅ
{ak } = {1, −1, −1, 1} áÅÐÊÑҳú¡Ç¹ {wk } = {−0.41, −0.33, 0.41, −0.59,
−1.29}
¨§¶Í´ÃËÑÊ¢éÍÁÙÅ
yk
´éÇÂǧ¨ÃµÃǨËÒ NPML àÁ× Íãªé
5.1) ǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 1 á·ç» ¤×Í
P (D) = −0.2613D
5.2) ǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 2 á·ç» ¤×Í
P (D) = −0.3929D − 0.5035D2
5.3) ǧ¨Ã¡Ãͧ·Ó¹ÒÂẺ 2 á·ç» ¤×Í
P (D) = −0.3443D − 0.4656D2 + 0.0964D3
6. ·ÓàËÁ×͹㹢éÍ 5 áµè¡Ó˹´ãËé·ÒÃìà¡çµ·Õ µéͧ¡Òà ¤×Í ÍÔ¹¾Øµ
{ak } = {−1, 1, 1, −1}
−1.12, −3.18}
áÅÐÅӴѺ¢éͼԴ¾ÅÒ´
H(D) = 1 + 2D + D2 ,
ÅӴѺ¢éÍÁÙÅ
{wk } = {0.47, 0.25, −0.38, −0.09,
º··Õ 7
ǧ¨ÃµÃǨËÒ PDNP
㹺·¹Õ ¨Ð¡ÅèÒǶ֧·Õ ÁÒ, ËÅÑ¡¡Ò÷ӧҹ, áÅлÃÐ⪹ì¢Í§ ǧ¨ÃµÃǨËÒ PDNP (pattern dependent noise predictive) [54] «Ö § à» ¹ ǧ¨ÃµÃǨËÒ·Õ ¶Ù¡ Í͡ẺÁÒà¾× Í ¨Ñ´¡ÒáѺ ÊÑҳú¡Ç¹¨ÔµàµÍÃì ¢Í§Ê× Í ºÑ¹·Ö¡ (media jitter noise) «Ö § ¾ººèÍÂã¹Ãкº¡Òúѹ·Ö¡ áÁèàËÅç¡ ·Õ ¤ÇÒÁ¨Ø ¢éÍÁÙÅ ÊÙ§æ ´Ñ§·Õ ¨Ð͸ԺÒµèÍä»ã¹º·¹Õ ¾ÃéÍÁ·Ñ § áÊ´§¼Å¡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾ÃÐËÇèҧǧ¨ÃµÃǨËÒ PDNP áÅÐǧ¨ÃµÃǨËÒ PRML
7.1
º·¹Ó
¨Ò¡·Õ ä´é͸ԺÒÂä»ã¹º··Õ 4 à·¤¹Ô¤ PRML ¤×Í à·¤¹Ô¤¡ÒÃãªé§Ò¹ÃèÇÁ¡Ñ¹ÃÐËÇèÒ§ÍÕ¤ÇÍäÅà«ÍÃìẺ PR (partial response equalizer) áÅÐǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ (Viterbi detector) «Ö §à» ¹·Õ ¹ÔÂÁãªé§Ò¹ ¡Ñ¹ ÁÒ¡ã¹Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì [27] à·¤¹Ô¤ PRML ¨Ð·Ó§Ò¹à» ¹ 2 ¢Ñ ¹µÍ¹ ¤×Í ¢Ñ ¹µÍ¹áá¨Ð·Ó¡ÒûÃѺ ÃÙ»ÃèÒ§¢Í§ÊÑÒ³·Õ ä´é ÃѺ ãËé à» ¹ 仵ÒÁÃÙ»ÃèÒ§¢Í§·ÒÃìà¡çµ (target) ·Õ µéͧ¡Òà áÅÐ¢Ñ ¹µÍ¹·Õ Êͧ¨Ð·Ó¡ÒöʹÃËÑÊ¢éÍÁÙÅâ´Âǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ·Õ ÊÃéÒ§¢Ö ¹¨Ò¡ ·ÒÃìà¡çµ·Õ ¡Ó˹´äÇé ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¶×Í ÇèÒ à» ¹ ǧ¨ÃµÃǨËÒÅӴѺ àËÁÒÐ·Õ ÊØ´ (optimal sequence detector) ¡çµèÍàÁ× Í Í§¤ì»ÃСͺ¢Í§ÊÑҳú¡Ç¹·Õ ὧÍÂÙèã¹ÊÑÒ³·Õ ¨Ð·Ó¡ÒöʹÃËÑÊ¢éÍÁÙÅÁÕÅѡɳР137
138
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
à» ¹ ÊÑҳú¡Ç¹à¡ÒÊì ÊÕ ¢ÒÇẺºÇ¡ [15] ÍÂèÒ§äáçµÒÁã¹·Ò§»¯ÔºÑµÔ ¡ÒÃãªé §Ò¹ÍÕ¤ÇÍäÅà«ÍÃì Ẻ PR ¨ÐÊ觼ŷÓãËéͧ¤ì»ÃСͺ¢Í§ÊÑҳú¡Ç¹·Õ ´éÒ¹¢Òà¢éҢͧǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔÁդس ÅѡɳÐà» ¹ ÊÑҳú¡Ç¹áººÊÕ (colored noise) â´Â੾ÒÐÍÂèÒ§ÂÔ § àÁ× Í ¤ÇÒÁ¨Ø ¢éÍÁÙÅ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ÊÙ§ (ND ÊÙ§) «Ö § ã¹¡Ã³Õ ¹Õ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨ÐäÁè ¶×Í ÇèÒ à» ¹ ǧ¨ÃµÃǨËÒÅӴѺ àËÁÒÐ ·Õ ÊØ´ÍÕ¡µèÍä» ´Ñ§¹Ñ ¹ ǧ¨ÃµÃǨËÒ NPML (noise predictive maximum likelihood) [51, 52] ¨Ö§ ä´é¶Ù¡¹ÓÁÒãªéà¾× Íà¾Ô Á»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ãкº â´Â·Õ ǧ¨ÃµÃǨËÒ NPML ¨ÐÁÕ¡Ãкǹ¡ÒÃ㹡ÒÃ·Ó ãËéÊÑҳú¡Ç¹à» ¹ÊÕ¢ÒÇ ¡è͹·Õ ¨ÐÊ觼ÅÅѾ¸ì·Õ ä´éä»·Ó¡ÒöʹÃËÑÊ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ã¹ º··Õ 6 áÊ´§ãËéàËç¹ÇèÒ Ç§¨ÃµÃǨËÒ NPML ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒǧ¨ÃµÃǨËÒ PRML â´Â੾ÒÐ ÍÂèÒ§ÂÔ §·Õ ¤ÇÒÁ¨Ø¢éÍÁÙŢͧÎÒÃì´´ÔÊ¡ìä´Ã¿ìÊÙ§ ¹Í¡¨Ò¡¹Õ ·Õ ND ÊÙ§æ ͧ¤ì»ÃСͺ¢Í§ÊÑҳú¡Ç¹·Õ ´éÒ¹¢Òà¢éÒ ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ Âѧ¨ÐÁÕÅÑ¡É³Ð¢Ö ¹ÍÂÙè¡Ñº Ẻ¢éÍÁÙÅ (data pattern) µÑÇÍÂèÒ§àªè¹ ÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× Í ºÑ¹·Ö¡ ¶×Íä´éÇèÒà» ¹ ÊÑҳú¡Ç¹·Õ ¢Ö ¹ÍÂÙè¡Ñºáºº¢éÍÁÙÅ (pattern dependent noise) ¡ÅèÒǤ×Í ÃдѺ ¤ÇÒÁÃعáç¢Í§ÊÑҳú¡Ç¹¨ÔµàµÍÃì ¢Í§Ê× Í ºÑ¹·Ö¡ ¨Ð¢Ö ¹ ÍÂÙè ¡Ñº Ẻ¢éÍÁÙÅ ·Õ à¢Õ¹ŧä»ã¹ Ê× ÍºÑ¹·Ö¡ à·¤¹Ô¤µèÒ§æ ä´é¶Ù¡¹ÓàʹÍà¾× Í·ÓãËéÊÑҳú¡Ç¹»ÃÐàÀ·¹Õ [54, 55] ÁÕÅѡɳСÅÒ ໠¹ÊÑҳú¡Ç¹ÊÕ¢ÒÇ (white noise) ¡è͹·Õ ¨Ð·Õ ·Ó¡ÒöʹÃËÑÊ¢éÍÁÙÅ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ã¹ÊèǹµèÍä»¹Õ ¨Ð͸ԺÒ¶֧ ËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒ PDNP [54] áÅÐà·¤¹Ô¤ ¡ÒÃÅ´¤ÇÒÁ «Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒ PDNP
7.2
¡ÒÃ¢Ö ¹ÍÂÙè¡Ñºà຺¢éÍÁÙŢͧÊÑҳú¡Ç¹
¾Ô¨ÒóÒẺ¨ÓÅͧªèͧÊÑÒ³ã¹ÃÙ»·Õ 7.1 ¡Ó˹´ãËé ¢Í§ ªèͧ ÊÑÒ³ ã¹ â´àÁ¹ µÍºÊ¹Í§ÃÇÁã¹â´àÁ¹
D
D, H(D)
C(D) = (1 − D)G(D)
¤×Í ¼Å µÍºÊ¹Í§ ·ÒÃìà¡çµã¹ â´àÁ¹
¤×Í ¼ÅµÍºÊ¹Í§
D, Q(D)
¤×Í ¼Å
¢Í§Ç§¨Ã¡Ãͧ¼èÒ¹µ Ó áÅÐÍÕ¤ÇÍäÅà«ÍÃì, áÅÐà¾× ÍãËé §èÒµèÍ ¡ÒÃ͸ԺÒÂ
ÍÕ¤ÇÍäÅà«ÍÃìẺ zero forcing [2, 16] ¨Ð¶Ù¡¹ÓÁÒãªéã¹Êèǹ¹Õ ¹Ñ ¹¤×Í
Q(D) = H(D)/C(D)
ã¹
·Ò§»¯ÔºÑµÔ ·ÒÃìà¡çµáÅÐÍÕ¤ÇÍäÅà«ÍÃì«Ö §¶Ù¡ÊÃéÒ§â´Âǧ¨Ã¡ÃͧàªÔ§àÊé¹áºº¼ÅµÍºÊ¹Í§ÍÔÁ¾ÑÅÊì¨Ó¡Ñ´ (FIR: nite impulse response) ¨ÐÁÕ¿ §¡ìªÑ¹¶èÒÂâ͹·Õ ᵡµèҧ仨ҡ ã¹·Ò§·ÄÉ®Õ à¾ÃÒÐ©Ð¹Ñ ¹ ¶éÒ ¡Ó˹´ãËé
H 0 (D)
áÅÐ
H(D)
áÅÐ
Q0 (D) = H 0 (D)/C(D)
Q(D)
·Õ µéͧ¡ÒÃ
¤×Í ¿ §¡ìªÑ¹ ¶èÒÂ
7.2.
¡ÒÃ¢Ö ¹ÍÂÙè¡Ñºà຺¢éÍÁÙŢͧÊÑҳú¡Ç¹
139
n(t)
ak {0,1}
1–D
bk
g(t)
sk
s(t)
p(t)
∆t k
LPF
tk = kT Q(D)
C(D)
yk
equalizer
âk
detector
H(D) ÃÙ»·Õ 7.1: Ẻ¨ÓÅͧªèͧÊÑÒ³
â͹¨ÃÔ§¢Í§·ÒÃìà¡çµáÅТͧÍÕ¤ÇÍäÅà«ÍÃì µÒÁÅӴѺ â´Â·Õ
H 0 (D) 6= H(D)
áÅÐ
Q0 (D) 6= Q(D)
´Ñ§¹Ñ ¹ ¢éÍÁÙÅàÍÒµì¾Øµ¢Í§ÍÕ¤ÇÍäÅà«ÍÃìÊÒÁÒöà¢Õ¹໠¹ÊÁ¡Ò÷ҧ¤³ÔµÈÒʵÃìä´é ¤×Í
Y (D) = A(D)C(D)Q0 (D) + N (D)Q0 (D) = A(D)H 0 (D) + N (D)Q0 (D) = A(D)H 0 (D) + N (D)Q0 (D) + [A(D)H(D) − A(D)H(D)] = A(D)H(D) + A(D)[H 0 (D) − H(D)] + N (D)Q0 (D) | {z }
(7.1)
W (D) àÁ× Í
N (D)
¤×Í ÊÑҳú¡Ç¹à¡ÒÊì ÊÕ ¢ÒÇẺºÇ¡ (AWGN) áÅÐ
W (D)
¤×Í ÊÑҳú¡Ç¹
W (D)
ÃÇÁ·Õ ä´é ³ ´éÒ¹¢Òà¢éҢͧǧ¨ÃµÃǨËÒÅӴѺ µÒÁÊÁ¡Òà (7.1) ¨ÐàËç¹ä´éÇèÒ ÊÑҳú¡Ç¹ ÁÕÅÑ¡É³Ð¢Ö ¹ÍÂÙè¡Ñºáºº¢éÍÁÙÅ (data dependent) à¹× ͧ¨Ò¡ ÁÕ¾¨¹ì
A(D)
ÍÂÙè
ÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡ à» ¹¼ÅÁÒ¨Ò¡ ¡ÒÃàÅ× Í¹µÓá˹觢ͧ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð áººÊØèÁ (random transition shift) ÃÐËÇèÒ§¡Ãкǹ¡Ò÷ÓãËé Ê× Í ºÑ¹·Ö¡ ÁÕ ÊÀÒ¾¤ÇÒÁà» ¹ áÁèàËÅç¡ (magnetization) µÒÁ·Õ µéͧ¡Òà â´Â¨ÐÁÕ ¿ §¡ìªÑ¹ ¤ÇÒÁ˹Òá¹è¹ ¤ÇÒÁ¹èÒ¨Ðà» ¹ Ẻà¡ÒÊì à«Õ¹·Õ ÁÕ ¤èÒ à©ÅÕ Â à·èÒ ¡Ñº ¤èÒ ÈÙ¹Âì áÅФèÒ ¤ÇÒÁá»Ã»Ãǹà·èÒ ¡Ñº ¨Ó¡Ñ´ãËéÁÕ¤èÒäÁèà¡Ô¹ ¤èÒÊÑÁºÙóì¢Í§
bk
T /2)
â´Â·Õ
σj
|bk |σj2
(¹Ñ ¹¤×Í
∆tk ∼ N (0, |bk |σj2 ))
¨Ð¶Ù¡¡Ó˹´à» ¹¨Ó¹Ç¹à»ÍÃìà«ç¹µì¢Í§ºÔµà«ÅÅì
T
à¹× ͧ¨Ò¡ µÓá˹觡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð¶Ù¡¡Ó˹´â´Â¢éÍÁÙźԵÍÔ¹¾Øµ
¤ÇÒÁÃعáç¢Í§ÊÑҳú¡Ç¹¨ÔµàµÍÃì ¢Í§Ê× Í ºÑ¹·Ö¡ ¨Ö§ ¢Ö ¹ ÍÂÙè ¡Ñº Ẻ¢éÍÁÙÅ ¢Í§
áÅÐ
áÅж١
|bk |
{ak }
{ak }
¤×Í
´Ñ§¹Ñ ¹
ÃÙ» ·Õ 7.2
140
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
5 4.5
Noise power (dB)
4 3.5 3 2.5 2 1.5 1 001
100
010
110
001
101
011
111
Data pattern
ÃÙ»·Õ 7.2: ¡ÓÅѧ ÊÑҳú¡Ç¹·Õ ¢Ö ¹ ¡Ñº Ẻ¢éÍÁÙÅ ³ ´éÒ¹¢ÒÍÍ¡¢Í§ÍÕ¤ÇÍäÅà«ÍÃì ·Õ ¶Ù¡ Í͡Ẻ ÊÓËÃѺ·ÒÃìà¡çµ EEPR2 [1 4 6 4 1] ¢Í§Ãкº¡Òúѹ·Ö¡áººá¹ÇµÑ § ·Õ ND = 2.5, SNR = 30 dB, áÅÐ
σj /T = 10%
áÊ´§¡ÓÅѧÊÑҳú¡Ç¹ (noise power) ·Õ ¢Ö ¹¡Ñºáºº¢éÍÁÙÅ ³ ´éÒ¹¢ÒÍÍ¡¢Í§ÍÕ¤ÇÍäÅà«ÍÃìẺ 21 á·ç» (tap) ·Õ ¶Ù¡Í͡ẺãËéÊÍ´¤Åéͧ¡Ñº·ÒÃìà¡çµ EEPR2,
H(D) = 1 + 4D + 6D2 + 4D3 +
D4 ,
¢Í§Ãкº¡Òúѹ·Ö¡ Ẻá¹ÇµÑ § (perpendicular recording) ·Õ ND = 2.5, SNR = 30 dB,
áÅÐ
σj /T = 10%
¨ÐàËç¹ä´éªÑ´à¨¹ÇèÒ ¡ÓÅѧÊÑҳú¡Ç¹¨ÐÁÕ¤èÒÊÙ§ àÁ× ÍÅӴѺ¢éÍÁÙÅÁÕ¡ÒÃà»ÅÕ Â¹
ʶҹÐËÅÒ¤ÃÑ § àªè¹ Ẻ¢éÍÁÙÅ 010 áÅÐ 101 à» ¹µé¹
7.3
ÍÑÅ¡ÍÃÔ·ÖÁ PDNP
¾Ô¨ÒóÒǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ [15] â´Â·Õ àÁµÃÔ¡ÊÒ¢Ò (branch metric) ¤Ó¹Ç³ä´é¨Ò¡
λk (u, v) = |yk −
ν X i=0
|
hi ak−1 |2 {z
r̂k (u,v)
}
(7.2)
7.3.
àÁ× Í
ÍÑÅ¡ÍÃÔ·ÖÁ PDNP
(u, v)
gram),
yk
141
á·¹¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð¨Ò¡Ê¶Ò¹Ð
u
ä»ÂѧʶҹÐ
v
ã¹á¼¹ÀÒ¾à·ÃÅÅÔÊ (trellis dia
¤×Í ¢éÍÁÙÅ·Õ ¨Ð·Ó¡ÒöʹÃËÑÊ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ,
ÊÑÒ³·Õ äÁè ÁÕ ÊÑҳú¡Ç¹·Õ ÊÍ´¤Åéͧ¡Ñº
(u, v)
r̂k (u, v)
¤×Í ¢éÍÁÙÅàÍÒµì¾Øµªèͧ
(¹Ñ ¹¤×Í ¤èÒ ·Õ áÊ´§ÍÂÙè ã¹áµèÅÐàÊé¹ ÊҢҢͧ
á¼¹ÀÒ¾à·ÃÅÅÔÊ àªè¹ µÒÁ·Õ áÊ´§ã¹ÃÙ»·Õ 4.6) «Ö §ËÒä´é¨Ò¡
rk = ak ∗ hk =
ν X
ak−i hi
(7.3)
i=0 àÁ× Í
∗
¤×Í µÑÇ´Óà¹Ô¹¡Òä͹âÇÅ٪ѹ (convolution operator),
µéͧ¡ÒÃ,
hk
¤×Í ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô µÑÇ·Õ
k
¢Í§·ÒÃìà¡çµ, áÅÐ
ν
H(D)
=
Pν
k=0 hk D
k ¤×Í ·ÒÃìà¡çµ·Õ
¤×Í Ë¹èǤÇÒÁ¨Ó¢Í§·ÒÃìà¡çµ
à¹× ͧ¨Ò¡ ͧ¤ì»ÃСͺ¢Í§ÊÑҳú¡Ç¹·Õ ὧÍÂÙèã¹¢éÍÁÙÅ
yk
«Ö §ËÒä´é¨Ò¡ (´ÙẺ¨ÓÅͧ¡ÒÃ
Í͡Ẻ·ÒÃìà¡çµã¹ÃÙ»·Õ 3.2)
wk = yk − rk D
â´Â¼Å¡ÒÃá»Å§
¢éÍÁÙÅ ¶éÒ Êè§ ¢éÍÁÙÅ
¤×Í
yk
W (D)
(7.4)
µÒÁÊÁ¡Òà (7.1) ¨ÐÁÕÅѡɳÐà» ¹ÊÑҳú¡Ç¹·Õ ¢Ö ¹ÍÂÙè¡Ñºáºº
à¢éÒ ä»·Ó¡ÒöʹÃËÑÊ ´éÇÂǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ »ÃÐÊÔ·¸ÔÀÒ¾ã¹ÃÙ» ¢Í§ÍѵÃÒ
¢éͼԴ¾ÅÒ´ºÔµ (BER) ·Õ ä´é ÃѺ ¨ÐäÁè ´Õ à¾ÃÒÐ©Ð¹Ñ ¹ 㹡ÒÃ·Õ ¨Ðà¾Ô Á »ÃÐÊÔ·¸ÔÀÒ¾ÃÇÁ¢Í§Ãкº ¨Ð µéͧÁÕ ¡ÒùӡÃкǹ¡ÒÃ㹡Ò÷ÓãËé ÊÑҳú¡Ç¹à» ¹ ÊÕ ¢ÒÇ ·ÓãËé¢éÍÁÙÅ
wk
(noise whitening process) à¾× Í
ÁÕÅѡɳÐà» ¹ÊÑҳú¡Ç¹ÊÕ¢ÒÇ ¡è͹·Õ ¨ÐÊ觼ÅÅѾ¸ì·Õ ä´éä»·Ó¡ÒöʹÃËÑÊ´éÇÂ
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ´Ñ§¹Ñ ¹ 㹡ÒÃãªé à·¤¹Ô¤ ¡Ò÷ӹÒÂÊÑҳú¡Ç¹ (noise prediction) ÃèÇÁ ¡ÑºÇ§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨Ð·ÓãËéàÁµÃÔ¡ÊÒ¢Òã¹ÊÁ¡Òà (7.2) µéͧ¶Ù¡´Ñ´á»Å§à» ¹
λk (u, v) = |yk − r̂k (u, v) − ŵk |2 àÁ× Í
ŵk
(7.5)
¤×Í ÊÑҳú¡Ç¹·Õ ¶Ù¡·Ó¹Ò «Ö §ËÒä´é¨Ò¡
ŵk =
L X
pi wk−i
(7.6)
i=1 â´Â·Õ á·ç»,
P (D) = pi
PL
i=1 pi D
i ¤×Í Ç§¨Ã¡Ãͧ·Ó¹Ò (prediction lter) ÊÑҳú¡Ç¹áºº
¤×Í ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô µÑÇ·Õ
i
L
¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂÊÑҳú¡Ç¹, áÅТéͼԴ¾ÅÒ´·Õ à¡Ô´¨Ò¡
142
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¡Ò÷ӹÒÂ
ek
¤×Í
ek = wk − ŵk ËÃ×Í
wk =
L X
(7.7)
pi wk−i + ek
(7.8)
i=1 â´Âǧ¨Ã¡Ãͧ·Ó¹ÒÂÊÑҳú¡Ç¹
P (D)
·Õ ´Õ¨Ðµéͧ·ÓãËé¢éͼԴ¾ÅÒ´·Õ à¡Ô´¨Ò¡¡Ò÷ӹÒÂ
ek
ÁÕ
ÅѡɳÐà» ¹ÊÑҳú¡Ç¹à¡ÒÊìÊÕ¢ÒÇãËéÁÒ¡·Õ ÊØ´ àÁ× ÍÃкº·Ó§Ò¹·Õ ¤ÇÒÁ¨Ø¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ìÊÙ§æ ¨Ð·ÓãËéÃдѺ¤ÇÒÁÃعáç¢Í§ÊÑҳú¡Ç¹ ¨Ð¢Ö ¹ ÍÂÙè ¡Ñº Ẻ¢éÍÁÙÅ ´Ñ§¹Ñ ¹ ǧ¨Ã¡Ãͧ·Ó¹ÒÂÊÑҳú¡Ç¹·Õ àËÁÒÐÊØ´ ÍÂÙè ¡Ñº Ẻ¢éÍÁÙÅ ´éÇÂàªè¹¡Ñ¹ à¾ÃÒÐ©Ð¹Ñ ¹ 㹡ÒÃËÒ¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§
P (D)
P (D)
¡ç ¤ÇÃ·Õ ¨Ð¢Ö ¹
¨Ðµéͧ¨Ñ´ ÃÙ» ÊÁ¡ÒÃ
(7.8) ãËÁè ´Ñ§¹Õ
wk (a) =
L X
pi (a)wk−i (a) + ek (a)
(7.9)
i=1 àÁ× Í
a
ãªé᷹Ẻ¢éÍÁÙŵèÒ§æ ·Õ à» ¹ä»ä´é ÊÁ¡Òà (7.9) ÊÒÁÒöà¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§àÁ·ÃÔ¡«ìä´é ¤×Í
wk (a) = p(a)T w(a) + ek (a) â´Â·Õ
p(a)
·Ó¹ÒÂ, áÅÐ
=
[p1 (a), p2 (a), . . . , pL (a)]T
w(a)
=
¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§ àÇ¡àµÍÃì
w(a)T
(7.10)
¤×Í àÇ¡àµÍÃì á¹ÇµÑ § ¢Í§¤èÒ ÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§Ç§¨Ã¡Ãͧ
[wk−1 (a), wk−2 (a), . . . , wk−L (a)]T P (D)
ÊÒÁÒö¤Ó¹Ç³ËÒä´é â´Â¡Òäٳ ·Ñ § Êͧ¢éÒ§¢Í§ÊÁ¡Òà (7.10) ´éÇÂ
áÅéÇãÊèµÑÇ´Óà¹Ô¹¡ÒäèÒ¤Ò´ËÁÒ (expectation operator) «Ö §¨Ðä´é¼ÅÅѾ¸ìà» ¹
£ ¤ £ ¤ E wk (a)w(a)T = p(a)T E w(a)w(a)T = p(a)T R(a) àÁ× Í
£ ¤ E w(a)w(a)T ¤×Í àÁ·ÃÔ¡«ìÍѵÊËÊÑÁ¾Ñ¹¸ì (auto correlation matrix) ¢Í§ÊÑÒ³ £ ¤ wk (a), áÅÐ E ek (a)w(a)T = 0 µÒÁËÅÑ¡¡ÒÃàªÔ§µÑ §©Ò¡ (orthogonality principle)
R(a)
ú¡Ç¹
(7.11)
=
[25] á¡éÊÁ¡Òà (7.11) ¨Ðä´é¼ÅÅѾ¸ìà» ¹
p(a)T = E[wk (a)w(a)T ] R−1 (a)
(7.12)
7.3.
ÍÑÅ¡ÍÃÔ·ÖÁ PDNP
143
áÅФèÒ¤ÇÒÁá»Ã»Ãǹ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒ (predictor error variance) ¤×Í [54]
£ ¤ £ ¤ £ ¤T σp2 (a) = E wk (a)2 − E wk (a)w(a)T R−1 (a)E wk (a)w(a)T
(7.13)
ÊÁ¡Òà (7.12) áÅÐ (7.13) áÊ´§ãËéàËç¹ÇèÒ ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂáÅФèÒ¤ÇÒÁá»Ã»Ãǹ ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒÂ¨Ð¢Ö ¹ÍÂÙè¡ÑºªèǧàÇÅÒ
a
k
·Õ ¾Ô¨ÒÃ³Ò (³ ªèǧàÇÅÒ
k
Ë¹Ö §æ ¡çÍÒ¨¨ÐÁÕẺ¢éÍÁÙÅ
k
·Õ µèÒ§¡Ñ¹ ä´é) ÍÂèÒ§äáçµÒÁ ¤èÒ ·Ñ § 2 ¤èÒ ¹Õ ÊÒÁÒö·Õ ¨Ð·ÓãËé äÁè ¢Ö ¹ ¡Ñº ªèǧàÇÅÒ
¡çä´é ¶éÒ ÊÁÁØµÔ ãËé
Ãкºà» ¹áººÊ൪ѹà¹ÃÕ (stationary) [10, 26] ËÃ×ͶéÒà» ¹Ãкº·Õ äÁèà» ¹áººÊ൪ѹà¹ÃÕ ÍÑÅ¡ÍÃÔ·ÖÁ Ẻ»ÃѺµÑÇ (adaptive algorithm) [4, 10, 16] ÊÒÁÒö·Õ ¨Ð¶Ù¡¹ÓÁÒãªé㹡ÒûÃѺ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§ ǧ¨Ã¡Ãͧ·Ó¹Ò áÅФèÒ¤ÇÒÁá»Ã»Ãǹ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒÂä´é à¹× ͧ¨Ò¡ ¤èÒ ¤ÇÒÁá»Ã»Ãǹ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒ («Ö § à» ¹ ¼ÅÁÒ¨Ò¡¡Ãкǹ¡ÒÃ㹡Ò÷ÓãËé ÊÑҳú¡Ç¹à» ¹ ÊÕ ¢ÒÇ) ¢Ö ¹ ÍÂÙè ¡Ñº Ẻ¢éÍÁÙÅ µÒÁ·Õ áÊ´§ã¹ÊÁ¡Òà (7.13) ¤èÒ àÁµÃÔ¡ ÊҢҢͧ ǧ¨ÃµÃǨËÒ PDNP ¨Ðµéͧ¤Ó¹Ö§¶Ö§¡ÒÃ¢Ö ¹ÍÂÙè¡Ñºáºº¢éÍÁÙÅ´éÇ à¾ÃÒÐ©Ð¹Ñ ¹ àÁµÃÔ¡ÊҢҢͧǧ¨Ã µÃǨËÒ PDNP ÊÒÁÒöà¢Õ¹ä´éà» ¹
λk (u, v) = log (σp (u, v)) + àÁ× Í
σp2 (u, v)
à¡Õ ÂÇ¢éͧ¡Ñº
|yk − r̂k (u, v) − ŵk (u, v)|2 2σp2 (u, v)
¤×Í ¤èÒ¤ÇÒÁá»Ã»Ãǹ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒÂ·Õ ÊÍ´¤Åéͧ¡Ñº
(u, v),
¢éÍÁÙÅ·Õ à¡Õ ÂÇ¢éͧ¡Ñº
áÅÐ
ŵk (u, v)
(u, v)
(7.14)
(u, v)
¤×Í ÊÑҳú¡Ç¹·Õ ¶Ù¡·Ó¹ÒÂ·Õ ÊÍ´¤Åéͧ¡Ñº
áÅÐẺ¢éÍÁÙÅ·Õ
(u, v)
áÅÐẺ
«Ö §ËÒä´é¨Ò¡
ŵk (u, v) =
L X
pi (u, v){yk−i − r̂k−i (u, v)}
(7.15)
i=1 ¨Ò¡ÊÁ¡Òà (7.12) áÅÐ (7.13) ¨Ð¾ºÇèÒ Ç§¨Ã¡Ãͧ·Ó¹ÒÂÊÑÒ³ á»Ã»Ãǹ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒÂ
P (D)
áÅÐ
σp2
σp2
P (D) =
PL i
pi D i
áÅФèÒ¤ÇÒÁ
¨Ð¢Ö ¹ÍÂÙè¡Ñºáºº¢éÍÁÙÅã¹áµèÅÐàÊé¹ÊÒ¢Ò ´Ñ§¹Ñ ¹ ¤èÒ¾ÒÃÒÁÔàµÍÃì
·Õ ãªé㹡ÒäӹdzàÁµÃÔ¡ÊÒ¢Òã¹áµèÅÐàÊé¹ÊҢҢͧἹÀÒ¾à·ÃÅÅÔʨÐÁÕ¤èÒµèÒ§¡Ñ¹
µÒÁẺ¢éÍÁÙÅ·Õ ÊÍ´¤Åéͧ¡Ñº¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
(u, v)
¹Ñ ¹æ
ã¹·Ò§»¯ÔºÑµÔáÅéÇ ¤èÒÊÑÁ»ÃÐÊÔ·¸Ô ¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂÊÑҳú¡Ç¹
p(a)
¨Ð¢Ö ¹ÍÂÙè¡Ñº¢éÍÁÙÅ
ºÔµÀÒÂã¹Ë¹éÒµèÒ§àÅ× Í¹áºº¨Ó¡Ñ´ ( nite sliding window) [56] «Ö §à¢Õ¹᷹´éÇÂÊÑÅѡɳì
aW k−M
144
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ÊÓËÃѺ¤èҨӹǹàµçÁºÇ¡ ʶҹÐà·èҡѺ
M
áÅÐ
2max(ν+L, M )+W
W
´Ñ§¹Ñ ¹ ǧ¨ÃµÃǨËÒ PDNP ¨Ðãªéá¼¹ÀÒ¾à·ÃÅÅÔÊ·Õ Áըӹǹ
ʶҹРÍÂèÒ§äáçµÒÁ à¾× ÍãËé§èÒµèÍ¡ÒÃ͸ԺÒÂËÅÑ¡¡Ò÷ӧҹ¢Í§
ÍÑÅ¡ÍÃÔ·ÖÁ PDNP ã¹Êèǹ¹Õ ¨Ð¾Ô¨ÒóÒ੾ÒСóշÕ
W =0
Ê¶Ò¹Ð·Ñ §ËÁ´·Õ ãªéã¹á¼¹ÀÒ¾à·ÃÅÅÔÊÁըӹǹà·èҡѺ
7.4
áÅÐ
M < ν +L
à¾ÃÒÐ©Ð¹Ñ ¹ ¨Ó¹Ç¹
2ν+L
ÍÑÅ¡ÍÃÔ·ÖÁ PS PDNP
ÍÑÅ¡ÍÃÔ·ÖÁ PDNP ·Õ ͸ԺÒÂä»ã¹ËÑÇ¢éÍ ·Õ 7.3 µéͧ¡ÒÃá¼¹ÀÒ¾à·ÃÅÅÔÊ ·Õ ÁÕ ¨Ó¹Ç¹Ê¶Ò¹Ð·Ñ §ËÁ´
2ν+L
ÍÂèÒ§äáçµÒÁ ¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒ PDNP ÊÒÁÒö·ÓãËéŴŧä´éâ´Âãªéá¹Ç¤Ô´ ¡ÒÃ
» ͹¡ÅѺ¤èҵѴÊԹ㨠(decision feedback) [53] áÅÐà¾× ÍãËéÃкºÁÕ»ÃÐÊÔ·¸ÔÀÒ¾·Õ ÂÍÁÃѺä´é á¹Ç¤Ô´ ¹Õ µéͧ¡ÒäÇÒÁÅÖ¡ (depth) ¡Òû ͹¡ÅѺ ¤èÒ µÑ´ÊÔ¹ã¨·Õ ¤è͹¢éÒ§¹éÍ àÁ× Í à·Õº¡Ñº ¤ÇÒÁÂÒÇ (ËÃ×Í ¨Ó¹Ç¹á·ç») ¢Í§Ç§¨Ã¡Ãͧ·Ó¹ÒÂÊÑҳú¡Ç¹ «Ö §ËÁÒ¤ÇÒÁÇèÒ á¼¹ÀÒ¾à·ÃÅÅÔÊ·Õ ãªéÂѧ¨Óà» ¹ µéͧ¶Ù¡ ¢ÂÒÂãËé ãËè ¢Ö ¹ (trellis expansion) ¹Ñ ¹¤×Í ÁÕըӹǹʶҹÐÁÒ¡¢Ö ¹ ¡ÇèÒ Ãкº·Õ äÁè ãªé ǧ¨Ã ¡Ãͧ·Ó¹ÒÂÊÑҳú¡Ç¹ ã¹Êèǹ¹Õ ¨Ð͸ԺÒ ÇÔ¸Õ¡ÒÃ·Õ àʹÍã¹ [53]
«Ö §¤ÅéÒ¡Ѻá¹Ç¤Ô´
¡ÒûÃÐÁÇżÅẺà¾Íà«ÍÃì
äÇàÇÍÃì (PSP: per survivor processing) [57] ÊÓËÃѺŴ¤ÇÒÁ«Ñº«é͹¢Í§ÍÑÅ¡ÍÃÔ·ÖÁ PDNP â´Â ÍÑÅ¡ÍÃÔ·ÖÁãËÁè·Õ ä´é ã¹Ë¹Ñ§Ê×Í¹Õ ¨ÐàÃÕ¡ÇèÒ ÍÑÅ¡ÍÃÔ·ÖÁ PS PDNP (per survivor PDNP algorith m) ǧ¨ÃµÃǨËÒ PS PDNP ¨Ð·Ó§Ò¹µÒÁÍÑÅ¡ÍÃÔ·ÖÁ PDNP º¹¾× ¹°Ò¹¢Í§á¹Ç¤Ô´ PSP «Ö § ¨Ð ·ÓãËé á¼¹ÀÒ¾à·ÃÅÅÔÊ ·Õ ãªé ã¹Ç§¨ÃµÃǨËÒ PS PDNP ÁÕ ¨Ó¹Ç¹Ê¶Ò¹Ðà·èÒ à´ÔÁ (ËÃ×Í à·èÒ ¡Ñº Ãкº·Õ äÁèãªèǧ¨Ã¡Ãͧ·Ó¹ÒÂÊÑҳú¡Ç¹) ¡ÅèÒǤ×Í á·¹·Õ ¨Ð·Ó¡ÒâÂÒÂá¼¹ÀÒ¾à·ÃÅÅÔÊãËéãËè¢Ö ¹ ǧ¨ÃµÃǨËÒ PS PDNP ¨Ð·Ó¡ÒÃÁͧÂé͹¡ÅѺ 仵ÒÁàÊé¹·Ò§·Õ Âѧ ÁÕ ªÕÇÔµ ÍÂÙè µÒÁá¼¹ÀÒ¾à·ÃÅÅÔÊ ·Õ ÁÒ¶Ö§ ³ ¨Ø´µèÍ (node) ·Õ ¡ÓÅѧ ¾Ô¨ÒÃ³Ò áÅéÇ ãªé ¢éÍÁÙÅ µèÒ§æ ·Õ ÊÍ´¤Åéͧ¡Ñº àÊé¹·Ò§·Õ Âѧ ÁÕ ªÕÇÔµ ÍÂÙè ¹Ñ ¹ 㹡ÒäӹdzËÒ¤èÒÊÑҳú¡Ç¹·Õ ¶Ù¡·Ó¹Ò ËÃ×ÍÍÒ¨¨Ð¡ÅèÒÇä´éÇèÒ Ç§¨ÃµÃǨËÒ PS PDNP Âѧ¤§ãªéÊÁ¡Òà (7.14) 㹡ÒäӹdzËÒ¤èÒàÁµÃÔ¡ÊÒ¢Ò Â¡àÇé¹áµè ÊÑҳú¡Ç¹·Õ ¶Ù¡·Ó¹Ò¨ж١ ¤Ó¹Ç³¨Ò¡
ŵk (u, v) =
L X i=1
pi (u, v)ẑk−i (u, v)
(7.16)
7.4.
ÍÑÅ¡ÍÃÔ·ÖÁ PS PDNP
time k (0) -1 -1
0
145
zˆk −2 ( A )
k+1
zˆk −1 ( A ) time k
zˆk ( 0,1)
2
k+1
A (1) 1 -1
(1)
0
zˆk −2 ( B )
2 -2
(2) -1 1
0 -2
ak = 1
0
ak = -1
(3) 1 1
B
zˆk ( 2,1) zˆk −1 ( B )
yk
yk −1 = 1
yk − 2 = 0.5
(a) Trellis diagram: PR4
yk = 0.15
(b) Decoding procedure
ÃÙ»·Õ 7.3: (a) á¼¹ÀÒ¾à·ÃÅÅÔÊÊÓËÃѺ·ÒÃìà¡çµáºº PR4 áÅÐ (b) ¢Ñ ¹µÍ¹¡ÒöʹÃËÑÊ¢éÍÁÙÅ´éÇ ἹÀÒ¾à·ÃÅÅÔÊ
àÁ× Í
ẑk−i (u, v)
¤×Í ÊÑҳú¡Ç¹·Õ ὧÍÂÙè ã¹¢éÍÁÙÅ
yk
·Õ à» ¹ ¼Å·ÓãËé à¡Ô´ ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
ÊÍ´¤Åéͧ¡ÑºàÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè·Õ à» ¹¼Å·ÓãËéà¡Ô´¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
(u, v)
«Ö §¹ÔÂÒÁâ´Â
ẑk (u, v) = yk (u, v) − r̂k (u, v)
(7.17)
´Ñ§¹Ñ ¹¨ÐàËç¹ä´éÇèÒ Ç§¨ÃµÃǨËÒ PS PDNP ¨Ðãªéá¼¹ÀÒ¾à·ÃÅÅÔÊ·Õ ÁըӹǹʶҹÐà·èҡѺ áµè¨ÐµéͧÁÕ¢Ñ ¹µÍ¹à¾Ô ÁàµÔÁ㹡ÒÃà¡çº¤èҢͧ
{ẑk−1 , ẑk−2 , . . . , ẑk−L }
2ν
ʶҹÐ
ÊÓËÃѺ·Ø¡àÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµ
ÍÂÙè
µÑÇÍÂèÒ§·Õ 7.1
ÃÙ»·Õ 7.3 áÊ´§µÑÇÍÂèÒ§¡ÒÃËÒ¤èÒÊÑҳú¡Ç¹·Õ ὧÍÂÙèã¹¢éÍÁÙÅ
ãËé à¡Ô´ ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹ÐÊÍ´¤Åéͧ¡Ñº
H(D) = 1 − D2 , Áըӹǹ
L=2
(u, v)
yk
·Õ à» ¹¼Å·Ó
µÒÁÊÁ¡Òà (7.17) ¢Í§Ãкº·Õ ãªé ·ÒÃìà¡çµ Ẻ PR4,
«Ö §ÁÕá¼¹ÀÒ¾à·ÃÅÅÔʵÒÁÃÙ»·Õ 7.3(a) àÁ× Íǧ¨Ã¡Ãͧ·Ó¹ÒÂÊÑҳú¡Ç¹·Õ ãªé
á·ç»
146
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ÇÔ¸Õ·Ó
ãËé¾Ô¨ÒóÒʶҹР(1) ³ àÇÅÒ
k+1
µÒÁ·Õ áÊ´§ã¹ÃÙ»·Õ 7.3(b) ¨ÐàËç¹ä´éÇèÒÁÕàÊé¹·Ò§·Õ ÇÔ §
à¢éÒÁÒËÒʶҹР(1) à» ¹¨Ó¹Ç¹ 2 àÊé¹·Ò§ ¤×Í àÊé¹·Ò§ A áÅÐ B à¾ÃÒÐ©Ð¹Ñ ¹ ¤èÒ ÊÁ¡Òà (7.16) ÊÓËÃѺ
i
ẑk−i (u, v)
ã¹
= 1 áÅÐ 2 ËÒä´é´Ñ§µèÍä»¹Õ ÊÓËÃѺàÊé¹·Ò§ A ¨Ðä´éÇèÒ
ẑk (0, 1) = yk − r̂k (0, 1) = 0.15 − 2 = −1.85 ẑk−1 (A) = ẑk−1 (0, 1) = yk−1 − r̂k−1 (0, 0) = 1 − 0 = 1 ẑk−2 (A) = ẑk−2 (0, 1) = yk−2 − r̂k−2 (0, 0) = 0.5 − 0 = 0.5 áÅÐÊÓËÃѺàÊé¹·Ò§ B ¨Ðä´éÇèÒ
ẑk (2, 1) = yk − r̂k (2, 1) = 0.15 − 0 = 0.15 ẑk−1 (B) = ẑk−1 (2, 1) = yk−1 − r̂k−1 (3, 2) = 1 − (−2) = 3 ẑk−2 (B) = ẑk−2 (2, 1) = yk−2 − r̂k−2 (1, 3) = 0.5 − 2 = −1.5
7.5
¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒ PDNP
㹡ÒÃà»ÃÕºà·Õº¤ÇÒÁ«Ñº«é͹ (complexity) ¢Í§Ç§¨ÃµÃǨËÒ ¨Ð¾Ô¨ÒóҨҡ¨Ó¹Ç¹µÑÇ´Óà¹Ô¹¡Òà (operator) ÊÓËÃѺ¡Òúǡ (addition) áÅСÒäٳ (multiplication) ·Õ µéͧãªé㹡Ò÷ӧҹ¢Í§áµèÅРǧ¨ÃµÃǨËÒ µÒÃÒ§·Õ 7.1 à»ÃÕºà·Õº¤ÇÒÁ«Ñº«é͹¢Í§Ç§¨ÃµÃǨËÒ PDNP áÅÐ PS PDNP àÁ× Í
Np
¤×Í ¨Ó¹Ç¹áºº¢éÍÁÙÅ·Õ ãªéã¹Ç§¨ÃµÃǨËÒ PDNP áÅÐ PS PDNP ¨Ò¡µÒÃÒ§¨ÐàËç¹ä´éÇèÒ Ç§¨Ã
µÃǨËÒ PS PDNP ÁÕ¤ÇÒÁ«Ñº«é͹áÅеéͧ¡ÒÃãªé˹èǤÇÒÁ¨Ó (memory requirement) ¹éÍ¡ÇèÒ Ç§¨ÃµÃǨËÒ PDNP ÁÒ¡
7.6
¼Å¡Ò÷´Åͧ
ã¹Êèǹ¹Õ ¨Ð·Ó¡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§Ç§¨ÃµÃǨËÒ PRML áÅÐ PDNP â´ÂãªéẺ¨ÓÅͧ ªèͧÊÑÒ³ µÒÁÃÙ»·Õ 7.1 àÁ× ÍÊÑÒ³ read back ÊÒÁÒöà¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§ÊÁ¡Òä³ÔµÈÒʵÃì
7.6.
¼Å¡Ò÷´Åͧ
147
µÒÃÒ§·Õ 7.1: ¨Ó¹Ç¹¢Í§µÑÇ´Óà¹Ô¹¡ÒÃ·Õ ãªéµèÍ¢éÍÁÙÅ 1 ºÔµ ¢Í§Ç§¨ÃµÃǨËÒ PDNP áÅÐ PS PDNP
ǧ¨ÃµÃǨËÒ
¨Ó¹Ç¹¢Í§µÑÇ´Óà¹Ô¹¡ÒÃ·Õ ãªéµèÍ¢éÍÁÙÅ 1 ºÔµ
(detector) ǧ¨ÃµÃǨËÒ PDNP
˹èǤÇÒÁ¨Ó·Õ µéͧ¡ÒÃ
¡Òúǡ
¡Òäٳ
(memory requirement)
(4L + 7)2ν+L
(2L + 8)2ν+L
(2L + 4)2ν+L + Np L + 2
(2L + 8)2ν
(2L + 8)2ν
(2L + 8)2ν + Np L
ǧ¨ÃµÃǨËÒ PS PDNP
ä´é¤×Í
p(t) =
S−1 X
bk g(t − kT + ∆tk ) + n(t)
(7.18)
k=0
ak ∈ {0, 1}
àÁ× Í
¤×Í ¢éÍÁÙÅ ºÔµ ÍÔ¹¾Øµ ·Õ ÁÕ ¨Ó¹Ç¹·Ñ §ËÁ´
ºÔµà»ÅÕ Â¹Ê¶Ò¹Ð (bk
= ±1
ÁÕ¡ÒÃà»ÅÕ Â¹á»Å§Ê¶Ò¹Ð),
S = 4096
ºÔµ,
bk = (ak − ak−1 )
ÊÍ´¤Åéͧ¡Ñº¡ÒÃà»ÅÕ Â¹á»Å§Ê¶Ò¹ÐºÇ¡ËÃ×Íź áÅÐ
g(t)
bk = 0
n(t)
¤×Í ÊÑÒ³
ú¡Ç¹à¡ÒÊìÊÕ¢ÒÇẺºÇ¡ (AWGN) ·Õ ÁÕ¤ÇÒÁ˹Òá¹è¹Ê໡µÃÑÁ¡ÓÅѧẺÊͧ´éÒ¹à·èҡѺ
∆tk
¤×ÍäÁè
¤×Í ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹ÐµÒÁÊÁ¡Òà (1.1) ÊÓËÃѺÃкº¡ÒÃ
ºÑ¹·Ö¡áººá¹Ç¹Í¹ áÅеÒÁÊÁ¡Òà (1.2) ÊÓËÃѺÃкº¡Òúѹ·Ö¡áººá¹ÇµÑ §,
áÅÐ
¤×Í
N0 /2,
¤×Í ÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡ (media jitter noise) ·Õ ÁÕ¿ §¡ìªÑ¹¤ÇÒÁ˹Òá¹è¹
¤ÇÒÁ¹èÒ¨Ðà» ¹ Ẻà¡ÒÊì à«Õ¹ (Gaussian probability density function) â´ÂÁÕ ¤èÒà©ÅÕ Â à·èÒ ¡Ñº ¤èÒ ÈÙ¹Âì áÅФèÒ ¤ÇÒÁá»Ã»Ãǹà·èÒ ¡Ñº à¡Ô¹
T /2)
àÁ× Í
σj
|bk |σj2
(¹Ñ ¹¤×Í
∆tk ∼ N (0, |bk |σj2 ))
¨Ð¶Ù¡ ¡Ó˹´à» ¹ ¨Ó¹Ç¹à»ÍÃìà«ç¹µì ¢Í§ºÔµ à«ÅÅì
(absolute value) ¢Í§
T
áÅж١ ¨Ó¡Ñ´ ãËé ÁÕ ¤èÒ äÁè
áÅÐ
|bk |
¤×Í ¤èÒÊÑÁºÙóì
bk
ÊÑÒ³ read back
p(t)
¨Ð¶Ù¡Ê觼èÒ¹ä»Âѧǧ¨Ã¡Ãͧ¼èÒ¹µ Ó (LPF: low pass lter) ºÑµà·ÍÃì
àÇÔÃìµÍѹ´Ñº·Õ 7 áÅж١·Ó¡ÒêѡµÑÇÍÂèÒ§´éǤÇÒÁ¶Õ ¡ÒêѡµÑÇÍÂèÒ§à·èҡѺ
1/T
â´ÂÊÁÁصÔÇèÒ¡Ãкǹ
¡ÒÃ㹡Òêѡ µÑÇÍÂèÒ§ÁÕ ¡ÒÃà¢éÒ ¨Ñ§ËÇÐÃÐËÇèÒ§ÊÑÒ³ read back áÅÐǧ¨ÃªÑ¡ µÑÇÍÂèҧẺÊÁºÙóì (perfect synchronization) ¨Ò¡¹Ñ ¹ ÅӴѺ¢éÍÁÙÅàÍÒµì¾Øµ
{sk } ¨Ð¶Ù¡» ͹ä»ÂѧÍÕ¤ÇÍäÅà«ÍÃì à¾× Í»ÃѺ
¤Ø³ÅѡɳТͧÊÑÒ³ãËéà» ¹ä»µÒÁ·ÒÃìà¡çµ·Õ µéͧ¡Òà áÅéÇ¡çÊè§ÅӴѺ¢éÍÁÙÅàÍÒµì¾Øµ
{yk }
·Ó¡ÒöʹÃËÑÊ¢éÍÁÙÅ´éÇÂǧ¨ÃµÃǨËÒ (detector) à¾× ÍËÒ¤èÒ»ÃÐÁÒ³¢Í§ÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ
·Õ ä´éä»
{ak }
·Õ
148
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
à» ¹ä»ä´éÁÒ¡·Õ ÊØ´ 㹺·¹Õ ¤èÒ SNR ¨Ð¹ÔÂÒÁâ´Â
µ SNR = 10 log10 àÁ× Í
Ei
Ei N0
¶ (7.19)
1
¤×Í ¾Åѧ§Ò¹¢Í§¼ÅµÍºÊ¹Í§ÍÔÁ¾ÑÅÊì¢Í§ªèͧÊÑÒ³
¹Í¡¨Ò¡¹Õ áµèÅШش ¢Í§ BER
¨Ð¶Ù¡ ¤Ó¹Ç³â´Âãªé ¢éÍÁÙÅ ËÅÒÂæ à«¡àµÍÃì (sector) ¨¹¡ÇèÒ ¨Ðä´é ¢éͼԴ¾ÅÒ´ºÔµ ÁÒ¡¡ÇèÒ ËÃ×Í à·èÒ ¡Ñº 1000 ºÔµ ǧ¨ÃµÃǨËÒ 3 Ẻ ¤×Í Ç§¨ÃµÃǨËÒ PRML, ǧ¨ÃµÃǨËÒ PDNP, áÅÐǧ¨ÃµÃǨËÒ PS PDNP ¨Ð¶Ù¡·Ó¡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾ ÊÓËÃѺÃкº·Õ ãªé·ÒÃìà¡çµáºº GPR3 (·ÒÃìà¡çµáºº 3 á·ç» ·Õ ¶Ù¡Í͡ẺµÒÁà§× ͹䢺ѧ¤ÑºáººâÁ¹Ô¡ [19]) áÅÐÍÕ¤ÇÍäÅà«ÍÃìẺ 21 á·ç» â´Â·Õ ·ÒÃìà¡çµ Ẻ GPR3 ÊÓËÃѺÃкº¡Òúѹ·Ö¡áººá¹Ç¹Í¹ ¤×Í Ãкº¡Òúѹ·Ö¡áººá¹ÇµÑ § ¤×Í
H(D) = 1 + 0.05D − 0.65D2
H(D) = 1+1.25D +0.62D2
áÅÐÊÓËÃѺ
ÃÙ»·Õ 7.4 à»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾
ã¹ÃÙ»¢Í§ BER ¢Í§Ç§¨ÃµÃǨËÒ·Ñ § 3 Ẻ ÊÓËÃѺÃкº¡Òúѹ·Ö¡áººá¹Ç¹Í¹áÅÐẺá¹ÇµÑ § ·Õ
N D = 2.5
áÅÐÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡
σj /T = 10%
¨Ò¡ÃÙ»·Õ 7.4(a) ÊÓËÃѺÃкº¡Òúѹ·Ö¡áººá¹Ç¹Í¹ ǧ¨ÃµÃǨËÒ PS PDNP ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ ã¡Åéà¤Õ§¡ÑºÇ§¨ÃµÃǨËÒ PDNP áµèǧ¨ÃµÃǨËÒ·Ñ §Êͧ¹Õ ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒǧ¨ÃµÃǨËÒ PRM L â´Â੾ÒÐÍÂèÒ§ÂÔ § ·Õ SNR ÊÙ§ (¹Ñ ¹¤×Í àÁ× Í ÊÑҳú¡Ç¹ËÅÑ¡ ã¹Ãкº ¤×Í ÊÑҳú¡Ç¹ ¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡) ã¹¢³Ð·Õ ÊÓËÃѺÃкº¡Òúѹ·Ö¡áººá¹ÇµÑ § (´ÙÃÙ»·Õ 7.4(b)) ǧ¨ÃµÃǨËÒ PRML ´ÙàËÁ×͹¨ÐÁÕ»ÃÐÊÔ·¸ÔÀÒ¾´Õ¡ÇèÒǧ¨ÃµÃǨËÒ PS PDNP áÅÐǧ¨ÃµÃǨËÒ PDNP àÅ硹éÍ ·Õ SNR µ Ó ·Ñ §¹Õ ÍÒ¨¨Ðà» ¹à¾ÃÒÐÇèÒ ·Õ SNR µ Ó ÊÑҳú¡Ç¹ËÅÑ¡ã¹ÃкºäÁèãªèÊÑҳú¡Ç¹ ¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡ ´Ñ§¹Ñ ¹ ǧ¨ÃµÃǨËÒ PS PDNP áÅÐǧ¨ÃµÃǨËÒ PDNP «Ö §¶Ù¡Í͡ẺÁÒãËé ¨Ñ´¡ÒáѺ ÊÑҳú¡Ç¹¨ÔµàµÍÃì ¢Í§Ê× Í ºÑ¹·Ö¡ ¨Ö§ äÁè ÊÒÁÒö·Ó§Ò¹ä´é ÍÂèÒ§ÁÕ »ÃÐÊÔ·¸ÔÀÒ¾ ÍÂèÒ§äà ¡çµÒÁ ·Õ SNR ÊÙ§ ǧ¨ÃµÃǨËÒ PS PDNP áÅÐǧ¨ÃµÃǨËÒ PDNP ¨ÐÁÕ »ÃÐÊÔ·¸ÔÀÒ¾´Õ ¡ÇèÒ Ç§¨Ã µÃǨËÒ PRML ÁÒ¡ ¨Ò¡¡Òüŷ´ÅͧáÊ´§ãËéàËç¹ä´éÇèÒ Ç§¨ÃµÃǨËÒ PS PDNP ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ ã¡Åéà¤Õ§¡Ñº ǧ¨ÃµÃǨËÒ PDNP à¾ÃÒÐ©Ð¹Ñ ¹ ã¹·Ò§»¯ÔºÑµÔ ǧ¨ÃµÃǨËÒ PS PDNP ÍÒ¨¨Ð¶Ù¡ ¹ÓÁÒãªé᷹ǧ¨ÃµÃǨËÒ PDNP à¹× ͧ¨Ò¡ ÁÕ¤ÇÒÁ«Ñº«é͹¹éÍ¡ÇèÒÁÒ¡ 1
¼ÅµÍºÊ¹Í§ÍÔÁ¾ÑÅÊì¢Í§ªèͧÊÑÒ³ ÁÕ¤èÒà·èҡѺ ͹ؾѹ¸ì¢Í§¼ÅµÍºÊ¹Í§¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
¡Òúѹ·Ö¡áººá¹Ç¹Í¹ áÅÐÁÕ¤èÒà·èҡѺ
0
g (t)/2
ÊÓËÃѺÃкº¡Òúѹ·Ö¡áººá¹ÇµÑ §
g 0 (t)
ÊÓËÃѺÃкº
7.7.
ÊÃØ»·éÒº·
149
−1
10
PRML (4 states) PS − PDNP (4 states) PDNP (32 states) −2
BER
10
−3
10
−4
10
−5
10
18
20
22
24
26
28
30
32
(a) SNR (dB) 0
10
PRML (4 states) PS − PDNP (4 states) PDNP (32 states) −1
BER
10
−2
10
−3
10
−4
10
18
20
22
24
26
28
30
32
34
36
(b) SNR (dB)
ÃÙ»·Õ 7.4:
»ÃÐÊÔ·¸ÔÀÒ¾ã¹ÃÙ» ¢Í§ BER ¢Í§Ç§¨ÃµÃǨËÒµèÒ§æ ÊÓËÃѺ Ãкº¡Òúѹ·Ö¡ (a) Ẻ
á¹Ç¹Í¹ áÅÐ (b) Ẻá¹ÇµÑ § ·Õ ND = 2.5 áÅÐ
7.7
σj /T = 10%
ÊÃØ»·éÒº·
àÁ× Í ¤ÇÒÁ¨Ø ¢éÍÁÙÅ ¢Í§ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì ÊÙ§ ¢Ö ¹ ÊÑҳú¡Ç¹·Õ ´éÒ¹¢Òà¢éÒ ¢Í§Ç§¨ÃµÃǨËÒÊÑÅÑ¡É³ì ¹Í¡¨Ò¡¨ÐÁÕÅѡɳÐà» ¹ÊÑҳú¡Ç¹áººÊÕ (colored noise) áÅéÇ Âѧ¨ÐÁÕÅÑ¡É³Ð¢Ö ¹ÍÂÙè¡Ñºáºº
150
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¢éÍÁÙÅ (data pattern) ´éÇ àªè¹ ÊÑҳú¡Ç¹¨ÔµàµÍÃì ¢Í§Ê× Í ºÑ¹·Ö¡ (media jitter noise) â´Â ·Õ ÃдѺ¤ÇÒÁÃعáç¢Í§ÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡¨Ð¢Ö ¹ÍÂÙè¡Ñºáºº¢éÍÁÙÅ·Õ à¢Õ¹ŧä»ã¹ Ê× Í ºÑ¹·Ö¡ ǧ¨ÃµÃǨËÒ PDNP ¨Ö§ ä´é ¶Ù¡ Í͡Ẻ¢Ö ¹ ÁÒ à¾× Í ¨Ñ´¡ÒáѺ ÊÑҳú¡Ç¹àËÅèÒ¹Õ â´Â ¨Ðà» ¹¡Ò÷ӧҹÃèÇÁ¡Ñ¹ÃÐËÇèҧǧ¨Ã¡Ãͧ·Ó¹ÒÂÊÑҳú¡Ç¹¡ÑºÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ ¶Ö§áÁéÇèÒ Ç§¨ÃµÃǨËÒ PDNP ¨ÐãËé »ÃÐÊÔ·¸ÔÀÒ¾·Õ ´Õ ¡ÇèÒ Ç§¨ÃµÃǨËÒ PRML ÁÒ¡ áµè ǧ¨Ã µÃǨËÒ PDNP ÁÕ ¤ÇÒÁ«Ñº«é͹ÁÒ¡¡ÇèÒ Ç§¨ÃµÃǨËÒ PRML à¾ÃÒÐÇèÒ á¼¹ÀÒ¾à·ÃÅÅÔÊ ·Õ ãªé 㹠ǧ¨ÃµÃǨËÒ PDNP ¨ÐÁÕ ¨Ó¹Ç¹Ê¶Ò¹Ðà¾Ô Á ÁÒ¡¢Ö ¹ à¹× ͧÁÒ¨Ò¡¨Ó¹Ç¹á·ç» ¢Í§Ç§¨Ã¡Ãͧ·Ó¹Ò ÊÑҳú¡Ç¹ » ËÒ¹Õ ÊÒÁÒö·Õ ¨Ðá¡éä¢ä´é â´Â ¡ÒûÃÐÂØ¡µì ãªé ÍÑÅ¡ÍÃÔ·ÖÁ PDNP µÒÁá¹Ç¤Ô´ ¢Í§ PSP «Ö §¨Ðä´é¼ÅÅѾ¸ìà» ¹ ǧ¨ÃµÃǨËÒ PS PDNP «Ö §ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ã¡Åéà¤Õ§¡ÑºÇ§¨ÃµÃǨËÒ PDNP áµèÁÕ¤ÇÒÁ«Ñº«é͹¹éÍ¡ÇèÒÁÒ¡
7.8
à຺½ ¡ËÑ´·éÒº·
1. ¨§Í¸ÔºÒÂ·Õ ÁҢͧá¹Ç¤Ô´¢Í§ÍÑÅ¡ÍÃÔ·ÖÁ PDNP
2. ¨§¾ÔÊÙ¨¹ì¤èÒ¤ÇÒÁá»Ã»Ãǹ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒ µÒÁÊÁ¡Òà (7.13)
3. ¨§¾ÔÊÙ¨¹ì¤èÒàÁµÃÔ¡ÊÒ¢Ò·Õ ãªéã¹Ç§¨ÃµÃǨËÒ PDNP µÒÁÊÁ¡Òà (7.14)
4. ¨§Í¸ÔºÒ¤ÇÒÁᵡµèÒ§¢Í§Ç§¨ÃµÃǨËÒ PRML, ǧ¨ÃµÃǨËÒ NPML, áÅÐǧ¨ÃµÃǨËÒ PDNP
5. ¨§Í¸ÔºÒÂËÅÑ¡¡Ò÷ӧҹ¢Í§Ç§¨ÃµÃǨËÒ PS PDNP
6. ¨§¾ÔÊÙ¨¹ì¨Ó¹Ç¹¢Í§µÑÇ´Óà¹Ô¹¡Òà (¡ÒúǡáÅСÒäٳ) ·Õ ãªéµèÍ¢éÍÁÙÅ 1 ºÔµ µÒÁ·Õ áÊ´§ã¹ µÒÃÒ§·Õ 7.1
7. ¨§¾ÔÊÙ¨¹ì¨Ó¹Ç¹Ë¹èǤÇÒÁ¨Ó·Õ µéͧ¡Òà µÒÁ·Õ áÊ´§ã¹µÒÃÒ§·Õ 7.1
º··Õ 8
¡ÒÃÍÍ¡à຺ÃËÑÊ RLL
㹺·¹Õ ¨Ð͸ԺÒ¶֧˹éÒ·Õ áÅÐËÅÑ¡¡Ò÷ӧҹ¢Í§ÃËÑÊ RLL (run length limited) [9] «Ö §à» ¹·Õ ¹ÔÂÁ ãªé§Ò¹ÁÒ¡ã¹Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì ¾ÃéÍÁ·Ñ §áÊ´§¢Ñ ¹µÍ¹¡ÒÃÍ͡Ẻ ÃËÑÊ RLL ÍÂèÒ§§èÒ à¾× Íãªé㹡ÒÃà¢éÒáÅжʹÃËÑÊ¢éÍÁÙÅ
8.1
º·¹Ó
ÃËÑÊ RLL ¤×Í ÃËÑÊÁÍ´ÙàŪѹ (modulation code) »ÃÐàÀ·Ë¹Ö §·Õ ¹ÔÂÁãªéÁÒ¡ã¹ÍØ»¡Ã³ìÎÒÃì´´ÔÊ¡ì ä´Ã¿ì â´Â¨Ð·Ó˹éÒ·Õ ã¹¡ÒáÓ˹´¨Ó¹Ç¹¢Í§ºÔµ 0 áÅкԵ 1 (µÒÁÃٻẺ¢Í§ NRZI) ·Õ àÃÕ§ µÔ´ ¡Ñ¹ ã¹ÅӴѺ ¢éÍÁÙÅ ·Õ µéͧ¡ÒèÐà¢Õ¹ŧä»ã¹Ê× Í ºÑ¹·Ö¡ â´Â·Ñ Çä» ÃËÑÊ RLL ¨Ð¶Ù¡ ¡Ó˹´´éÇ ¾ÒÃÒÁÔàµÍÃì 4 µÑÇ ¤×Í
1)
m
2)
n
m, n, d,
áÅÐ
k
â´Â¨ÐÍÂÙèã¹ÃÙ»¢Í§ÃËÑÊ
m/n (d, k)
àÁ× Í
¤×Í ¨Ó¹Ç¹¢éÍÁÙźԵÍÔ¹¾Øµ (µèÍ¡ÒÃà¢éÒÃËÑÊË¹Ö §¤ÃÑ §) ·Õ ¨Ð·Ó¡ÒÃà¢éÒÃËÑÊ RLL
¤×Í ¨Ó¹Ç¹¢éÍÁÙźԵàÍÒµì¾Øµ (µèÍ¡ÒÃà¢éÒÃËÑÊË¹Ö §¤ÃÑ §) ·Õ ä´é¨Ò¡¡ÒÃà¢éÒÃËÑÊ RLL â´Â·Ñ Çä»
n≥m
àÊÁÍ
3)
d
¤×Í àÅ¢¨Ó¹Ç¹àµçÁ·Õ ¡Ó˹´¨Ó¹Ç¹·Õ¹ éÍÂ·Õ ÊØ´¢Í§ºÔµ 0 ·Õ ÍÂÙèÃÐËÇèÒ§ºÔµ 1
4)
k
¤×Í àÅ¢¨Ó¹Ç¹àµçÁ·Õ ¡Ó˹´¨Ó¹Ç¹·ÕÁ Ò¡·Õ ÊØ´¢Í§ºÔµ 0 ·Õ ÍÂÙèÃÐËÇèÒ§ºÔµ 1 151
152
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
# input bits
RLL code
(m bits)
# output bits (n bits)
ÃÙ»·Õ 8.1: Ẻ¨ÓÅͧ¡ÒÃà¢éÒÃËÑÊ RLL
àÁ× Í ¢éÍÁÙźԵ 1 ¨ÐÊÍ´¤Åéͧ¡Ñº¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð (transition) ¢Í§¡ÃÐáÊä¿¿ Òà¢Õ¹ (write cur rent) ·Õ ¨Ð» ͹à¢éÒä»ã¹ËÑÇà¢Õ¹ (write head) à¾× Í·ÓãËéÊ× ÍºÑ¹·Ö¡ ³ ºÃÔàdz·Õ µéͧ¡ÒèÐà¢Õ¹¢éÍÁÙŠŧä»ÁÕÊÀÒ¾¤ÇÒÁà» ¹áÁèàËÅç¡ (magnetization) µÒÁ·Õ µéͧ¡Òà Êèǹ¢éÍÁÙźԵ 0 ËÁÒ¶֧ äÁèÁÕ¡Òà à»ÅÕ Â¹Ê¶Ò¹Ð¢Í§¡ÃÐáÊä¿¿ Òà¢Õ¹ à¾ÃÒÐ©Ð¹Ñ ¹ ¾ÒÃÒÁÔàµÍÃì
d
¨ÐªèÇ·ÓãËé ºÔµ 1 ÊͧºÔµ ÍÂÙèËèÒ§
¡Ñ¹ «Ö § ¨ÐªèÇÂÅ´¼Å¡Ãзº¢Í§¡ÒÃá·Ã¡ÊÍ´ÃÐËÇèÒ§ÊÑÅѡɳì (ISI: intersymbol interference) Êèǹ¾ÒÃÒÁÔàµÍÃì
k
¨ÐªèÇÂÃѺ»ÃСѹÇèÒ ÅӴѺ¢éÍÁÙÅ·Õ ¨Ðà¢Õ¹ŧä»ã¹Ê× ÍºÑ¹·Ö¡¨ÐÁÕºÔµà»ÅÕ Â¹Ê¶Ò¹Ð
à¡Ô´¢Ö ¹ÊÁ ÓàÊÁÍà¾Õ§¾Í à¾× Í·Õ ¨Ð·ÓãËéÃкºä·ÁÁÔ §ÃԤѿàÇÍÐÃÕ (timing recovery) ÊÒÁÒö·Ó§Ò¹ä´é ÍÂèÒ§ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ µÑÇÍÂèÒ§àªè¹ ¶éÒÅӴѺ¢éÍÁÙÅ·Õ ¨Ðà¢Õ¹ŧä»ã¹Ê× ÍºÑ¹·Ö¡ ¤×Í
···11111 ··· 11111··· ÅӴѺ ¢éÍÁÙÅ ¹Õ ¶×Í ÇèÒ à» ¹ ÅӴѺ ¢éÍÁÙÅ ·Õ äÁè ´Õ à¹× ͧ¨Ò¡ ¨Ð·ÓãËé à¡Ô´ » ËÒ ISI ÍÂèÒ§Ãعáç ã¹·Ò§ µÃ§¡Ñ¹¢éÒÁ ¶éÒÅӴѺ¢éÍÁÙÅ·Õ ¨Ðà¢Õ¹ŧä»ã¹Ê× ÍºÑ¹·Ö¡ ¤×Í
···00000 ··· 00000··· ¡ç ¨Ð¶×Í ÇèÒ à» ¹ ÅӴѺ ¢éÍÁÙÅ ·Õ äÁè ´Õ àªè¹¡Ñ¹ à¹× ͧ¨Ò¡ ¨Ð·ÓãËé à¡Ô´ » ËÒàÃ× Í§¡ÒÃà¢éÒ ¨Ñ§ËÇÐ (synchro nization) ¢Í§Ãкºä·ÁÁÔ § ÃÔ ¤Ñ¿àÇÍÐÃÕ à¾ÃÒÐ©Ð¹Ñ ¹ à¾× Í ËÅÕ¡àÅÕ Â§ÅӴѺ ¢éÍÁÙÅ ·Ñ § 2 Ẻ¹Õ ¨Ö§ ÁÕ ¤ÇÒÁ¨Óà» ¹ ·Õ ¨Ðµéͧà¢éÒ ÃËÑÊ ÅӴѺ ¢éÍÁÙÅ ´éÇÂÃËÑÊ RLL «Ö § ã¹·Ò§»¯ÔºÑµÔ áÅéÇ ¡ÒÃà¢éÒ áÅжʹÃËÑÊ ´éÇÂÃËÑÊ RLL ÊÒÁÒö·Óä´é§èÒÂâ´Â¡ÒÃãªé µÒÃÒ§¤é¹ËÒ (look up table) 㹡ÒÃà¢éÒáÅжʹÃËÑÊ ¢éÍÁÙÅ ÃÙ»·Õ 8.1 áÊ´§áºº¨ÓÅͧ¡ÒÃà¢éÒÃËÑÊ RLL â´Â·Õ ÍѵÃÒÃËÑÊ (code rate) ¨Ð¹ÔÂÒÁâ´Â ¨Ó¹Ç¹ ºÔµÍÔ¹¾Øµ
m
ËÒôéǨӹǹºÔµàÍÒµì¾Øµ
n
¹Ñ ¹¤×Í
R=
m ≤1 n
(8.1)
8.2.
¨Ó¹Ç¹ÅӴѺ¢éÍÁÙÅ·Ñ §ËÁ´·Õ ÊÍ´¤Åéͧ¡Ñºà§× ͹䢺ѧ¤Ñº
à¹× ͧ¨Ò¡ ¨Ó¹Ç¹ºÔµàÍÒµì¾Øµ
m
n
(D, K)
153
·Õ ä´é¨Ò¡¡ÒÃà¢éÒÃËÑʨÐÁըӹǹÁÒ¡¡ÇèÒËÃ×Íà·èҡѺ¨Ó¹Ç¹ºÔµÍÔ¹¾Øµ
àÊÁÍ ´Ñ§¹Ñ ¹ ¢éÍàÊÕ¢ͧ¡ÒÃà¢éÒÃËÑÊ RLL ·Õ àËç¹ä´éªÑ´à¨¹¡ç¤×Í ¨Ð·ÓãËéà¡Ô´ ºÔµÊèǹà¡Ô¹ (redun
dant bit) «Ö § ¨Ð·ÓãËé ÊÙàÊÕ à¹× Í·Õ ¡ÒèѴ à¡çº ¢éÍÁÙÅ ·Õ µéͧ¡ÒÃã¹ÎÒÃì´ ´ÔÊ¡ì ä´Ã¿ì 仺ҧÊèǹ ´Ñ§¹Ñ ¹ 㹡ÒÃàÅ×Í¡ÃËÑÊ RLL ã´ÁÒãªé §Ò¹ ¡ç ¤ÇÃ·Õ ¨ÐàÅ×Í¡ãªé ÃËÑÊ RLL ·Õ ÁÕ ÍѵÃÒÃËÑÊ
R
à¢éèÒã¡Åé ¤èÒ 1
ãËé ÁÒ¡·Õ ÊØ´ à¾× Í Å´¡ÒÃÊÙàÊÕ à¹× Í·Õ ¡ÒèѴ à¡çº ¢éÍÁÙÅ ·Õ µéͧ¡Òà ÊÓËÃѺ à¹× ÍËÒ㹺·¹Õ ¨Ð͸ԺÒ¶֧ ¢Ñ ¹µÍ¹¡ÒÃÍ͡ẺÃËÑÊ RLL ÍÂèÒ§§èÒ àÁ× Í¡Ó˹´¾ÒÃÒÁÔàµÍÃì
8.2
m, n, d,
áÅÐ
¨Ó¹Ç¹ÅӴѺ¢éÍÁÙÅ·Ñ §ËÁ´·Õ ÊÍ´¤Åéͧ¡Ñºà§× ͹䢺ѧ¤Ñº
¡Ó˹´ãËé ÅӴѺ ¢éÍÁÙÅ ÁÕ ¤ÇÒÁÂÒÇ·Ñ §ÊÔ ¹ ºÑ§¤Ñº (constraint)
(d, k)
L
k
ÁÒãËé
(d, k)
ºÔµ ¨Ó¹Ç¹ÅӴѺ ¢éÍÁÙÅ ·Ñ §ËÁ´ ·Õ ÊÍ´¤Åéͧ¡Ñº à§× ͹ä¢
ÊÒÁÒöËÒä´é¨Ò¡ÊÁ¡ÒõèÍä»¹Õ [9]
N (L) = L + 1, 1 ≤ L ≤ d + 1
(8.2)
N (L) = N (L − 1) + N (L − d − 1), d + 1 ≤ L ≤ k
(8.3)
N (L) = d + k + 1 − L +
k X
N (L − i − 1), k < L ≤ d + k
(8.4)
i=d
N (L) =
k X
N (L − i − 1), L > d + k
(8.5)
i=d
àÁ× Í
N (L) = 0
ÊÓËÃѺ
L<0
ã¹¡Ã³Õ ·Õ ¾ÒÃÒÁÔàµÍÃì
áÅÐ
k =∞
N (0) = 1
¨Ó¹Ç¹ÅӴѺ ¢éÍÁÙÅ ·Ñ §ËÁ´·Õ ÊÍ´¤Åéͧ¡Ñº à§× ͹䢺ѧ¤Ñº
(d, ∞)
¨ÐÊÒÁÒöËÒä´é¨Ò¡ÊÁ¡ÒõèÍ仹Õ
àÁ× Í
Nd (L) = 0
·Ñ §ËÁ´
Nd (L)
Nd (L) = L + 1, 1 ≤ L ≤ d + 1
(8.6)
Nd (L) = Nd (L − 1) + Nd (L − d − 1), L > d + 1
(8.7)
ÊÓËÃѺ
L<0
·Õ ÁÕ¤ÇÒÁÂÒÇ
L
áÅÐ
Nd (0) = 1
µÒÃÒ§·Õ 8.1 áÊ´§µÑÇÍÂèÒ§¨Ó¹Ç¹ÅӴѺ ¢éÍÁÙÅ
·Õ ÊÍ´¤Åéͧ¡Ñºà§× ͹䢺ѧ¤Ñº
(d, ∞)
154
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
µÒÃÒ§·Õ 8.1: ºÑ§¤Ñº
µÑÇÍÂèÒ§¨Ó¹Ç¹ÅӴѺ ¢éÍÁÙÅ ·Ñ §ËÁ´
Nd (L)
·Õ ÁÕ ¤ÇÒÁÂÒÇ
L
·Õ ÊÍ´¤Åéͧ¡Ñº à§× ͹ä¢
(d, ∞) d
L=4
L=5
L=6
L=7
L=8
L=9
L = 10
1
8
13
21
34
55
89
144
2
6
9
13
19
28
41
60
3
5
7
10
14
19
26
36
4
5
6
8
11
15
20
26
5
5
6
7
9
12
16
21
àÁ× Í·ÃÒº¨Ó¹Ç¹ÅӴѺ¢éÍÁÙÅ·Ñ §ËÁ´
K
·Ñ §ËÁ´
N (L)
·Õ ÊÍ´¤Åéͧ¡Ñºà§× ͹䢺ѧ¤Ñº
(d, k)
·Õ ÊÒÁÒö¹ÓÁÒãªéá·¹ÅӴѺ¢éÍÁÙÅáµèÅÐÅӴѺ¨ÐÁÕ¤èÒà·èҡѺ
K = d log2 {N (L)} e (bits) àÁ× Í
dxe
áÅéÇ ¨Ó¹Ç¹ºÔµ
(8.8)
x
á·¹¨Ó¹Ç¹àµçÁ ºÇ¡·Õ ¹éÍÂ·Õ ÊØ´ ·Õ ÁÕ ¤èÒ ÁÒ¡¡ÇèÒ ËÃ×Í à·èÒ ¡Ñº ¤èÒ
ÊÒÁÒöãªé¢éÍÁÙźԵ¨Ó¹Ç¹ 3 ºÔµ
{000,
àªè¹ ¶éÒ
N4 (6) = 8
¡ç
001, 010, 011, 100, 101, 110, 111} 㹡ÒÃá·¹ÅӴѺ
¢éÍÁÙÅáµèÅÐẺ
8.3
¤ÇÒÁ¨Ø¢Í§ÃËÑÊ RLL à຺
(d, k)
ã¹·ÄɯբͧÃкºÊ× ÍÊÒà ¤ÇÒÁ¨Ø (capacity) ËÁÒ¶֧ ¤èÒÊÙ§ÊØ´¢Í§ÍѵÃÒÃËÑÊ
R
·Õ ÊÒÁÒö·ÓãËé
ÊÑÁÄ·¸Ô¼Åä´é «Ö §¨Ð¹ÔÂÒÁâ´Â [9]
1 log2 {N (L)} L→∞ n
C(d, k) = lim àÁ× Í
N (L)
C(d, k)
¤×Í ¨Ó¹Ç¹ÅӴѺ¢éÍÁÙÅ·Ñ §ËÁ´·Õ ÊÍ´¤Åéͧ¡Ñºà§× ͹䢺ѧ¤Ñº
(8.9)
(d, k)
¹Í¡¨Ò¡¹Õ ¤èÒ¤ÇÒÁ¨Ø
Âѧ à» ¹ ¾ÒÃÒÁÔàµÍÃì ·Õ ºè§ºÍ¡¶Ö§ ¤ÇÒÁÊÒÁÒö㹡ÒèѴ à¡çº ¢éÍÁÙÅ ¢èÒÇÊÒâͧ¼Ùéãªé ·Õ µéͧ¡ÒÃ
¨Ðà¢Õ¹ŧä»ã¹Ê× Í ºÑ¹·Ö¡ (äÁè ¹Ñº ºÔµ Êèǹà¡Ô¹) ¹Ñ ¹¤×Í ¶éÒ ¤èÒ ÊÒÁÒö¨Ñ´à¡çº¢éÍÁÙÅ¢èÒÇÊÒâͧ¼Ùéãªéä´éÁÒ¡àªè¹¡Ñ¹
C(d, k)
ÂÔ § ÁÒ¡ ¡ç áÊ´§ÇèÒ Ãкº
8.3.
¤ÇÒÁ¨Ø¢Í§ÃËÑÊ RLL à຺
(D, K)
155
¹Í¡¨Ò¡¹Õ 㹡ÒÃà»ÃÕºà·Õº»ÃÐÊÔ·¸ÔÀÒ¾¢Í§ÃËÑÊ RLL ẺµèÒ§æ ÊÒÁÒö·Óä´éâ´Â¡ÒþԨÒÃ³Ò ·Õ ¤èÒ »ÃÐÊÔ·¸Ô¼Å¢Í§ÃËÑÊ (code e ciency) «Ö §¹ÔÂÒÁâ´Â
η=
R C(d, k)
(8.10)
¡ÅèÒǤ×Í ÃËÑÊ RLL ·Õ ÁÕ¤èÒ»ÃÐÊÔ·¸Ô¼Å¢Í§ÃËÑÊÁÒ¡ ¡çáÊ´§ÇèÒÁÕ»ÃÐÊÔ·¸ÔÀҾ㹡ÒÃãªé§Ò¹ÊÙ§
8.3.1
(d, k)
ÍѵÃÒ¢èÒÇÊÒÃàªÔ§àÊ鹡ӡѺ¢Í§ÃËÑÊ RLL à຺
à¹× ͧ¨Ò¡ ¡ÒäӹdzËÒ¤èÒ ¤ÇÒÁ¨Ø
C(d, k)
´Ñ§¹Ñ ¹ â´Â·Ñ Çä» ¨Ö§¹ÔÂÁ¤Ó¹Ç³ËÒ¤èÒ
L → ∞
ã¹ÊÁ¡Òà (8.9) àÁ× Í
C(d, k)
·Óä´é ¤è͹¢éÒ§ÅÓºÒ¡
¨Ò¡ ÍѵÃÒ¢èÒÇÊÒÃàªÔ§àÊ鹡ӡѺ (asymptotic infor
mation rate) «Ö §¹ÔÂÒÁâ´Â [58]
C(d, k) = log2 {λmax } àÁ× Í
λmax
¤×Í ÃÒ¡¨Ó¹Ç¹¨ÃÔ§ (real root) ·Õ ÁÕ¤èÒÁÒ¡ÊØ´ ¢Í§ÊÁ¡ÒÃ
xk+2 − xk+1 − xk−d+1 + 1 = 0,
k<∞
(8.12)
xd+1 − xd − 1 = 0,
k=∞
(8.13)
µÒÃÒ§·Õ 8.2 áÊ´§ÍѵÃÒ¢èÒÇÊÒÃàªÔ§ àÊé¹ ¡Ó¡Ñº
C(d, k)
¡ÒÃá¡éÊÁ¡Òà (8.11) (8.13) áÅÐàÁ× Í¾Ô¨ÒóҤèÒ ãªé·Ñ Ç仨ÐÁÕ¤èÒ¾ÒÃÒÁÔàµÍÃì
8.3.2
(8.11)
d≤2
¢Í§ÃËÑÊ RLL Ẻ
C(d, k)
(d, k)
µèÒ§æ ·Õ ä´é ¨Ò¡
ã¹µÒÃÒ§·Õ 8.2 ¨Ð¾ºÇèÒ ÃËÑÊ RLL ·Õ
àÊÁÍ à¾× Í·Õ ¨ÐÃѺ»ÃСѹä´éÇèÒÍѵÃÒÃËÑÊ
R ≥ 1/2
ÍѵÃÒ¤ÇÒÁ˹Òàà¹è¹
ÍѵÃÒ¤ÇÒÁ˹Òá¹è¹ DR (density ratio) ËÃ×Í ¤ÇÒÁ˹Òá¹è¹ ¡ÒúÃèØ
(packing density) à» ¹
¾ÒÃÒÁÔàµÍÃì ·Õ ºè§ºÍ¡¶Ö§ ÃÐÂзҧ·Ò§¡ÒÂÀÒ¾ (physical distance) ÃÐËÇèÒ§µÓáË¹è§ ¢Í§¡ÒÃà»ÅÕ Â¹ Ê¶Ò¹Ð·Õ µÔ´¡Ñ¹ 2 µÓáË¹è§ ¢Í§ÅӴѺ¢éÍÁÙÅ·Õ à¢éÒÃËÑÊ RLL «Ö §¹ÔÂÒÁâ´Â
DR = (1 + d)R
(8.14)
156
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
µÒÃÒ§·Õ 8.2: ÍѵÃÒ¢èÒÇÊÒÃàªÔ§àÊ鹡ӡѺ¢Í§ÃËÑÊ RLL Ẻ
â´Â·Õ
R
d=1
d=2
d=3
d=4
(d, k)
k
d=0
1
0.6942
2
0.8792
0.4057
3
0.9468
0.5515
0.2878
4
0.9752
0.6174
0.4057
0.2232
5
0.9881
0.6509
0.4650
0.3218
0.1823
10
0.9997
0.6909
0.5418
0.4460
0.3746
0.3158
15
0.9999
0.6939
0.5501
0.4615
0.3991
0.3513
∞
1.0000
0.6942
0.5515
0.4650
0.4057
0.3620
µèÒ§æ
d=5
¤×Í ÍѵÃÒÃËÑÊ
µÒÃÒ§·Õ 8.3 áÊ´§¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§¤ÇÒÁ¨Ø àÁ× Í¤ÇÒÁ¨Ø
C(d, k)
C(d, k) áÅÐÍѵÃÒ¤ÇÒÁ˹Òá¹è¹ DR ¨ÐàËç¹ä´éÇèÒ
Ŵŧ ¤èÒÍѵÃÒ¤ÇÒÁ˹Òá¹è¹ DR ¡ç¨Ðà¾Ô Á¢Ö ¹ ¤ÇÒÁÊÑÁ¾Ñ¹¸ì¹Õ ÊÒÁÒö͸ԺÒÂä´é
´Ñ§µèÍä»¹Õ ¨Ò¡ÊÁ¡Òà (8.14) àÁ× Í¾ÒÃÒÁÔàµÍÃì
d
à¾Ô Á¢Ö ¹ ¤èÒ DR ¡ç¨Ðà¾Ô Á¢Ö ¹ áµè¾ÒÃÒÁÔàµÍÃì
d
¢Ö ¹ ¹Õ ÁÕ ¤ÇÒÁËÁÒÂÇèÒ ¢éÍÁÙÅ ·Õ ¶Ù¡ à¢éÒ ÃËÑÊ ¨ÐÁÕ ºÔµ Êèǹà¡Ô¹ à¾Ô Á ÁÒ¡¢Ö ¹ (à¾ÃÒÐÇèÒ ¾ÒÃÒÁÔàµÍÃì
·Õ à¾Ô Á
d
¤×Í
¨Ó¹Ç¹ºÔµ 0 ¹éÍÂÊØ´ ·Õ ÍÂÙè ÃÐËÇèÒ§ºÔµ 1) ¶éÒ ¾Ô¨ÒóÒÇèÒ Ê× Í ºÑ¹·Ö¡ ÁÕ à¹× Í·Õ ã¹¡ÒèѴ à¡çº ¢éÍÁÙÅ ·Õ ¨Ó¡Ñ´ ´Ñ§¹Ñ ¹ ÃкºÊÒÁÒö·Õ ¨Ð¨Ñ´à¡çº¢éÍÁÙÅ¢èÒÇÊÒâͧ¼Ùéãªéä´é¹éÍÂŧ à¹× ͧ¨Ò¡ µéͧàËÅ×Íà¹× Í·Õ ºÒ§ÊèǹäÇé ÊÓËÃѺ¨Ñ´à¡çººÔµÊèǹà¡Ô¹ à¾ÃÒÐ©Ð¹Ñ ¹ ¨Ö§Ê觼ŷÓãËé¤èÒ¤ÇÒÁ¨Ø
8.4
C(d, k)
·Õ ¤Ó¹Ç³ä´éÁÕ¤èÒ¹éÍÂŧ
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ (FSM: nite state machine) ¢Í§ÃËÑÊ RLL ¨ÐáÊ´§ãËéàË繶֧ ¡ÒÃà»ÅÕ Â¹á»Å§ ¢Í§Ê¶Ò¹Ðã¹ÃËÑÊ RLL µÒÁà§× ͹䢺ѧ¤Ñº ¢Í§¾ÒÃÒÁÔàµÍÃì
(d, k)
¨ÐÁÕà¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´µÒÁÃÙ»·Õ 8.2 àÁ× Í
Si
(d, k)
¤×Í Ê¶Ò¹Ð
i
µÑÇÍÂèÒ§àªè¹ ÃËÑÊ RLL Ẻ
áÅеÑÇàÅ¢·Õ áÊ´§ÍÂÙèµÒÁàÊé¹ÅÙ¡ÈÃ
¤×Í ¢éÍÁÙźԵàÍÒµì¾Øµ·Õ ÊÍ´¤Åéͧ¡Ñºà§× ͹䢺ѧ¤Ñº¢Í§¾ÒÃÒÁÔàµÍÃì
(d, k) ¨Ò¡ÃÙ»·Õ 8.2 ʶҹÐàÃÔ Áµé¹
8.4.
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL
157
µÒÃÒ§·Õ 8.3: ¤ÇÒÁÊÑÁ¾Ñ¹¸ìÃÐËÇèÒ§¤ÇÒÁ¨Ø
0
S1
S2
0
C(d, k)
d
C(d, ∞)
1
0.6942
1.3884
2
0.5515
1.6545
3
0.4650
1.8600
4
0.4057
2.0285
5
0.3620
2.1720
DR =
0
Sd
áÅÐÍѵÃÒ¤ÇÒÁ˹Òá¹è¹ DR
(1 + d)C(d, ∞)
S d +1
0
Sk
1
1
ÃÙ»·Õ 8.2: à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ
¨ÐÍÂÙè·Õ ʶҹÐ
S1
S k +1
1
(d, k)
«Ö §ãËé¶×ÍÇèÒà» è¹à˵ءÒóì·Õ à¨ÍºÔµ 1 µÑÇááã¹ÅӴѺ¢éÍÁÙÅ à¾ÃÒÐ©Ð¹Ñ ¹ ºÔµµèÍä»
¨Ðµéͧ໠¹ºÔµ 0 à» ¹¨Ó¹Ç¹ÍÂèÒ§¹éÍ ä»Âѧ ʶҹÐ
0
Sd+1 )
d
µÑǵԴµè͡ѹ (¹Ñ ¹¤×Í Ê¶Ò¹Ð
¾ÍÅӴѺ ¢éÍÁÙÅ ÁÕ ºÔµ 0 ¤Ãº
d
S1
¡ç¨Ðà´Ô¹·Ò§à» ¹àÊ鹵ç
µÑÇ áÅéÇ ¨Ò¡à§× ͹䢺ѧ¤Ñº
(d, k)
áÊ´§ÇèÒ ºÔµ
µÑǶѴä»ÊÒÁÒö໠¹ä´é·Ñ §ºÔµ 0 ËÃ×ͺԵ 1 «Ö §¶éÒà» ¹ºÔµ 1 àÁ× Íã´ Ãкº¡ç¨ÐµéͧÇÔ §¡ÅѺä»àÃÔ Áµé¹·Õ ʶҹÐ
S1
ãËÁè áµè¶éÒà» ¹ºÔµ 0 ¡ç¨ÐÁÕºÔµ 0 ä´éÍÕ¡äÁèà¡Ô¹
k−d
µÑÇ áÅÐàÁ× ÍÁÕºÔµ 0 µÔ´µè͡ѹ¤Ãº
µÑÇáÅéÇ ºÔµµÑǶѴ仨еéͧ໠¹ºÔµ 1 à·èÒ¹Ñ ¹ ¹Ñ ¹¤×Í Ãкº¨Ð¶Ù¡ºÑ§¤ÑºãËé¡ÅѺä»àÃÔ Áµé¹·Õ ʶҹÐ
k
S1
ãËÁèâ´ÂÍѵâ¹ÁѵÔ
µÑÇÍÂèÒ§·Õ 8.1
¨§áÊ´§á¼¹ÀÒ¾à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL µÒÁà§× ͹䢺ѧ¤Ñº¢Í§¾ÒÃÒÁÔàµÍÃì
(d, k) = (1, 3) ÇÔ¸Õ·Ó
¾ÒÃÒÁÔàµÍÃì
(1, 3)
ËÁÒ¶֧ ÅӴѺ¢éÍÁÙŨÐÁÕºÔµ 0 ÍÂèÒ§¹éÍÂË¹Ö §µÑÇ ËÃ×ÍÍÂèÒ§ÁÒ¡ÊÒÁµÑÇ
158
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
S1
0
0
S2
0
S3
1
S4
1
1
ÃÙ»·Õ 8.3: à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ
(1, 3)
µÔ´µè͡ѹ ·Õ ÍÂÙèÃÐËÇèÒ§ºÔµ 1 «Ö §ÊÒÁÒöà¢Õ¹໠¹à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ä´é µÒÁÃÙ»·Õ 8.3
8.5
àÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ ¢Í§ÃËÑÊ RLL Ẻ
(d, k)
ÊÒÁÒöà¢Õ¹ãËé ÍÂÙè ã¹ÃÙ» ¢Í§ àÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹
ʶҹР(state transition matrix) ä´é «Ö §¹ÔÂÒÁâ´Â àÁ·ÃÔ¡«ì á¹ÇµÑ § â´Â·Õ ÊÁÒªÔ¡¢Í§àÁ·ÃÔ¡«ì
D(i, j)
¹Ñ ¹¤×Í á¶Ç·Õ
i
D
·Õ ÁÕ¢¹Ò´
áÅÐá¹ÇµÑ §·Õ
(k + 1) j
á¶Ç áÅÐ
¨Ð¶Ù¡¡Ó˹´â´Â
D(i, 1) = 1, i ≥ d + 1 1, j = i + 1 D(i, j) = 0, else µÑÇÍÂèÒ§àªè¹ à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ðä´é´Ñ§¹Õ
(1, 3)
(k + 1)
(8.15)
µÒÁÃÙ»·Õ 8.3 ÊÒÁÒöà¢Õ¹໠¹àÁ·ÃÔ¡«ì
0 1 0 0
1 0 1 0 D= 1 0 0 1 1 0 0 0 àÁ·ÃÔ¡«ì
D
(8.16)
ã¹ÊÁ¡Òà (8.16) ÊÒÁÒöÊÃéÒ§ä´é ´Ñ§µèÍä»¹Õ ¶éÒ ¡Ó˹´ãËé áµèÅÐá¶Ç᷹ʶҹÐ
áµèÅРʶҹР¡ÅèÒǤ×Í á¶Ç ·Õ Ë¹Ö § ãªé á·¹ ʶҹÐ
S1
áÅÐ á¶Ç ·Õ Êͧ ãªé á·¹ ʶҹÐ
S2
àªè¹à´ÕÂǡѹãËéáµèÅÐá¹ÇµÑ §á·¹Ê¶Ò¹ÐáµèÅÐʶҹР¡ÅèÒǤ×Í á¹ÇµÑ §·Õ Ë¹Ö §ãªé᷹ʶҹÐ
à» ¹µé¹
S1
áÅÐ
8.5.
àÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
159
S2
à» ¹µé¹ ´Ñ§¹Ñ ¹ 㹡ÒÃÊÃéÒ§àÁ·ÃÔ¡«ì ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð¨Ò¡à¤Ã× Í§
á¹ÇµÑ § ·Õ Êͧãªé ᷹ʶҹÐ
ʶҹШӡѴ¢Í§ÃËÑÊ RLL Ẻ (ʶҹÐ
S1 )
(1, 3)
µÒÁÃÙ»·Õ 8.3 ãËé¾Ô¨ÒóҷÕÅÐá¹ÇµÑ § àªè¹ ã¹á¹ÇµÑ §·Õ Ë¹Ö §
Si
ãËé ´Ù ÇèÒ ÁÕ àÊé¹ ÅÙ¡ÈèҡʶҹÐ
ÇèÒ ÁÕàÊé¹ÅÙ¡ÈèҡʶҹÐ
S2 , S3 ,
áÅÐ
S4
ã´ºéÒ§·Õ ÇÔ § à¢éÒ ÁÒ·Õ Ê¶Ò¹Ð
S1
¨Ò¡ÃÙ» ·Õ 8.3 ¨Ð¾º
´Ñ§¹Ñ ¹ ã¹á¹ÇµÑ §·Õ Ë¹Ö §¹Õ ¤èÒ 1 ¨Ð¶Ù¡ãÊèà¢éÒä»ã¹á¶Ç·Õ
Êͧ, á¶Ç·Õ ÊÒÁ, áÅÐá¶Ç·Õ ÊÕ Êèǹá¶Ç·Õ Ë¹Ö §¨ÐãËéà» ¹¤èÒ 0 㹷ӹͧà´ÕÂǡѹ ¶éÒ¾Ô¨ÒÃ³Ò·Õ á¹ÇµÑ § ·Õ Êͧ (ʶҹÐ
S2 )
ãËé´ÙÇèÒÁÕàÊé¹ÅÙ¡ÈèҡʶҹÐ
¾ºÇèÒ ÁÕàÊé¹ÅÙ¡ÈèҡʶҹÐ
S1
Si
ã´ºéÒ§·Õ ÇÔ §à¢éÒÁÒ·Õ Ê¶Ò¹Ð
àÊé¹à´ÕÂÇ·Õ ÇÔ §ÁÒ·Õ Ê¶Ò¹Ð
S2
S2
¨Ò¡ÃÙ»·Õ 8.3 ¨Ð
´Ñ§¹Ñ ¹ ¤èÒ 1 ¨Ð¶Ù¡ãÊèà¢éÒä»ã¹á¶Ç·Õ
Ë¹Ö § Êèǹá¶ÇÍ× ¹æ ¨ÐÁÕ¤èÒà» ¹¤èÒ 0 à» ¹µé¹ ÊÓËÃÑºã¹¡Ã³Õ·Õ ¾ÒÃÒÁÔàµÍÃì
k=∞
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ
à¢Õ¹ãËéÍÂÙèã¹ÃÙ»¢Í§àÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð ·Õ ÊÁÒªÔ¡¢Í§àÁ·ÃÔ¡«ì
D(i, j)
D
·Õ ÁÕ¢¹Ò´
(d + 1)
á¶Ç áÅÐ
(d, ∞)
(d + 1)
ÊÒÁÒö
á¹ÇµÑ § â´Â
¨Ð¶Ù¡¡Ó˹´â´Â
D(i, j) = 1, j = i + 1 D(d + 1, 1) = D(d + 1, d + 1) = 1 D(i, j) = 0, else
µÑÇÍÂèÒ§·Õ 8.2
¨§áÊ´§á¼¹ÀÒ¾à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ áÅÐàÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð ¢Í§ÃËÑÊ RLL
µÒÁà§× ͹䢺ѧ¤Ñº¢Í§¾ÒÃÒÁÔàµÍÃì
ÇÔ¸Õ ·Ó
(8.17)
¾ÒÃÒÁÔàµÍÃì
(0, 3)
(d, k) = (0, 3)
ËÁÒ¶֧ ÅӴѺ ¢éÍÁÙÅ ¨ÐÁÕ ºÔµ 0 ÍÂèÒ§¹éÍÂË¹Ö § µÑÇ ËÃ×Í ÍÂèÒ§ÁÒ¡ÊÒÁ
µÑÇ µÔ´µèÍ ¡Ñ¹ ·Õ ÍÂÙè ÃÐËÇèÒ§ ºÔµ 1 «Ö § ÊÒÁÒö à¢Õ¹ à» ¹ à¤Ã× Í§ ʶҹР¨Ó¡Ñ´ ä´é µÒÁ ÃÙ» ·Õ 8.4 â´Â ·Õ àÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð·Õ ÊÍ´¤Åéͧ¡Ñºà¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¹Õ ¤×Í
1 1 0 0
1 0 1 0 D= 1 0 0 1 1 0 0 0
160
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
1
0
S1
S2
0
0
S3
1
S4
1
1
ÃÙ»·Õ 8.4: à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ
8.5.1
(0, 3)
¡ÒÃËÒÍѵÃÒ¢èÒÇÊÒÃàªÔ§àÊ鹡ӡѺ
àÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
D
ÊÒÁÒö¹ÓÁÒãªé㹡ÒäӹdzËÒÍѵÃÒ¢èÒÇÊÒÃàªÔ§àÊ鹡ӡѺ
C(d, k)
¨Ò¡ÊÙµÃã¹ÊÁ¡Òà (8.11) ¹Ñ ¹¤×Í
© ª C(d, k) = log2 λD max â´Â·Õ
D λmax
ʶҹÐ
D
(8.18)
¤×Í ¤èÒÅѡɳÐ੾ÒШӹǹ¨ÃÔ§ (real eigenvalue) ·Õ ÁÕ¤èÒÁÒ¡ÊØ´¢Í§àÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹
«Ö §ËÒä´é¨Ò¡¡ÒÃá¡éÊÁ¡ÒÃ
det(D − λI) = 0 àÁ× Í det(·) ¤×Í ¡ÒÃËÒ¤èÒ´Õà·ÍÃìÁÔá¹¹µì (determinant), ·Õ ÁÕ¢¹Ò´à·èҡѺàÁ·ÃÔ¡«ì
8.5.2
D,
áÅÐ
λ
¤×Í àÁ·ÃÔ¡«ìàÍ¡Åѡɳì (identity matrix)
¤×Í ¤èÒÅѡɳÐ੾ÒÐ
ÅӴѺ¢éÍÁÙÅ·Õ ÊÍ´¤Åéͧ¡Ñºà¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL à຺
¨Ò¡à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ
(d, k)
¨Ó¹Ç¹ÅӴѺ¢éÍÁÙÅ·Õ à» ¹ä»ä´é·Ñ §ËÁ´·Õ ÁÕ¤ÇÒÁÂÒÇ
Sj
I
(8.19)
¨ÐÁÕ¤èÒà·èҡѺ ¤èҢͧÊÁÒªÔ¡á¶Ç·Õ
µÑÇÍÂèÒ§·Õ 8.3
i
·Õ áÊ´§ã¹ÃÙ»·Õ 8.2 àÁ× Í
L ºÔµ ·Õ ÍÍ¡¨Ò¡Ê¶Ò¹Ð Si
áÅÐá¹ÇµÑ §·Õ
j
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ
¢éÍÁÙÅ·Õ à» ¹ä»ä´é·Ñ §ËÁ´·Õ ÁÕ¤ÇÒÁÂÒÇ
2
Si
¢Í§àÁ·ÃÔ¡«ì
(0, 2)
ºÔµ ·Õ ÍÍ¡¨Ò¡Ê¶Ò¹Ð
DL
(d, k)
¤×Í Ê¶Ò¹Ð
i
´Ñ§¹Ñ ¹
áÅéÇä»ÊÔ ¹ÊØ´·Õ ʶҹÐ
¹Ñ ¹¤×Í
DL (i, j)
áÊ´§ã¹ÃÙ»·Õ 8.5 ¨§ËҨӹǹÅӴѺ
Si
áÅéÇä»ÊÔ ¹ÊØ´·Õ ʶҹÐ
Sj
ÊÓËÃѺ
8.5.
àÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
161
1 S1
0
0
S2
1
S3
1
ÃÙ»·Õ 8.5: à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ
(0, 2)
0 ≤ i, j ≤ 3 ÇÔ¸Õ ·Ó
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ ¢Í§ÃËÑÊ RLL Ẻ
àÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
D
(0, 2)
ã¹ÃÙ» ·Õ 8.5 ÊÒÁÒöà¢Õ¹ãËé ÍÂÙè ã¹ÃÙ» ¢Í§
ä´é ¤×Í
1 1 0
D= 1 0 1 1 0 0
2 ºÔµ 2 1 1 D2 = 2 1 0 1 1 0
´Ñ§¹Ñ ¹ ¨Ó¹Ç¹ÅӴѺ¢éÍÁÙÅ·Õ à» ¹ä»ä´é·Ñ §ËÁ´·Õ ÁÕ¤ÇÒÁÂÒÇ
(8.20)
ËÒä´é¨Ò¡
D2
«Ö §ÁÕ¤èÒà·èҡѺ
(8.21)
ÊÁ¡Òà (8.21) ºÍ¡ãËé·ÃÒºÇèÒ
ÅӴѺ¢éÍÁÙÅ·Õ ÍÍ¡¨Ò¡Ê¶Ò¹Ð
S1 → S1
Áըӹǹà·èҡѺ
D(1, 1) = 2
µÑÇ ¤×Í
{01,
11}
ÅӴѺ¢éÍÁÙÅ·Õ ÍÍ¡¨Ò¡Ê¶Ò¹Ð
S2 → S1
Áըӹǹà·èҡѺ
D(2, 1) = 2
µÑÇ ¤×Í
{01,
11}
ÅӴѺ¢éÍÁÙÅ·Õ ÍÍ¡¨Ò¡Ê¶Ò¹Ð
S3 → S1
Áըӹǹà·èҡѺ
D(3, 1) = 1
µÑÇ ¤×Í
{11}
ÅӴѺ¢éÍÁÙÅ·Õ ÍÍ¡¨Ò¡Ê¶Ò¹Ð
S1 → S2
Áըӹǹà·èҡѺ
D(1, 2) = 2
µÑÇ ¤×Í
{10}
162
8.6
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¢Ñ ¹µÍ¹¡ÒÃÍÍ¡à຺ÃËÑÊ RLL
ã¹Êèǹ¹Õ ¨ÐáÊ´§¢Ñ ¹µÍ¹¡ÒÃÍ͡ẺµÒÃÒ§¤é¹ËÒ à¾× Íãªé㹡ÒÃà¢éÒáÅжʹÃËÑÊ RLL â´ÂãªéµÑÇÍÂèÒ§ ´Ñ§µèÍ仹Õ
µÑÇÍÂèÒ§·Õ 8.4
¨§ÊÃéÒ§µÒÃÒ§¤é¹ËÒÊÓËÃѺ¡ÒÃà¢éÒáÅжʹÃËÑÊ RLL Ẻ
¤ÃÑ §·Õ ¼èÒ¹¡ÒÃà¢éÒÃËÑÊáÅéǨÐÁÕ¤ÇÒÁÂÒÇà·èҡѺ
ÇÔ¸Õ·Ó
n=3
(0, 2)
â´Â·Õ ¢éÍÁÙÅáµèÅÐ
¾ÃéÍÁ·Ñ §ËÒ»ÃÐÊÔ·¸Ô¼Å¢Í§ÃËÑÊ
η
¡ÒÃÊÃéÒ§µÒÃÒ§¤é¹ËÒÊÓËÃѺ¡ÒÃà¢éÒáÅжʹÃËÑÊ RLL ÊÒÁÒöáºè§ÍÍ¡à» ¹ 4 ¢Ñ ¹µÍ¹ ´Ñ§¹Õ
¢Ñ ¹µÍ¹·Õ 1:
ãËé¾Ô¨ÒóҴÙÇèÒ ¢éÍÁÙÅ 3 ºÔµÁÕ·Ñ §ËÁ´¡Õ Ẻ «Ö §¨Ðä´éÇèÒ ÁÕ·Ñ §ËÁ´ 8 Ẻ ¤×Í
{000, 001, 010, 011, 100, 101, 110, 111} ¢Ñ ¹µÍ¹·Õ 2:
ãËé¾Ô¨ÒóҴÙÇèÒ ÁÕ¢éÍÁÙŪشä˹ºéÒ§·Õ äÁèÊÍ´¤Åéͧ¡Ñºà§× ͹䢺ѧ¤Ñº
ÇèÒ ÁÕ¢éÍÁÙÅ·Ñ §ËÁ´ 7 Ẻ·Õ ¼èÒ¹à§× ͹䢺ѧ¤Ñº
(0, 2)
(0, 2)
«Ö §¨Ð¾º
¹Ñ ¹¤×Í
{001, 010, 011, 100, 101, 110, 111} ¢Ñ ¹µÍ¹·Õ 3:
ãËéÅͧ¹Ó¢éÍÁÙÅ·Õ ä´é¨Ò¡¢Ñ ¹µÍ¹·Õ 2 áµèÅеÑÇ ÁÒ·Ó¡Òõè͡ѹ·Ñ §·Ò§«éÒÂáÅзҧ¢ÇÒ
áÅéÇ´ÙÇèÒ ÁÕ¢éÍÁÙŵÑÇä˹ºéÒ§·Õ àÁ× Í¹ÓÁÒµè͡ѹáÅéÇ ¨ÐäÁèÊÍ´¤Åéͧ¡Ñºà§× ͹䢺ѧ¤Ñº ¾ºÇèÒ ¢éÍÁÙÅ 001 áÅÐ 100 àÁ× Í¹Ó仵è͡Ѻ¢éÍÁÙŵÑÇÍ× ¹¨Ð·ÓãËéà§× ͹䢺ѧ¤Ñº
(0, 2)
(0, 2)
ã¹·Õ ¹Õ ¨Ð
¼Ô´ä» ´Ñ§¹Ñ ¹
¢éÍÁÙÅ 001 áÅÐ 100 ¨Ðµéͧ¶Ù¡ µÑ´ ·Ô § ä» ·ÓãËé ¢éÍÁÙÅ ·Õ ËŧàËÅ×Í ÍÂÙè ÁÕ à¾Õ§ 5 Ẻ ·Õ ¼èÒ¹à§× Í¹ä¢ ºÑ§¤Ñº
(0, 2)
¹Ñ ¹¤×Í
{010, 011, 101, 110, 111} «Ö §¢éÍÁÙÅàËÅèÒ¹Õ ¡ç¤×Í ¢éÍÁÙÅ·Õ ÊÒÁÒö¹ÓÁÒãªéà» ¹¢éÍÁÙÅ·Õ ¼èÒ¹¡ÒÃà¢éÒÃËÑÊ RLL áÅéÇ ¢Ñ ¹µÍ¹·Õ 4:
¨Ò¡¢éÍÁÙÅ·Ñ § 5 Ẻ·Õ ä´éã¹¢Ñ ¹µÍ¹·Õ 3 ãËéàÅ×Í¡ÁÒ 4 Ẻ (Ẻ㴡çä´é) à¾× Íãªéã¹
¡ÒÃÊÃéÒ§µÒÃÒ§¤é¹ËÒ ÊÓËÃѺ¡ÒÃà¢éÒáÅжʹÃËÑÊ¢éÍÁÙÅÍÔ¹¾Øµ·ÕÅÐ 2 ºÔµ ¹Ñ ¹¤×Í 11} «Ö §¨Ðä´é µÒÁµÒÃÒ§·Õ 8.4 áÅШÐä´éÇèÒ ÍѵÃÒÃËÑÊ ¤×Í
{00,
01,
10,
8.6.
¢Ñ ¹µÍ¹¡ÒÃÍÍ¡à຺ÃËÑÊ RLL
163
µÒÃÒ§·Õ 8.4: µÒÃÒ§¤é¹ËÒÊÓËÃѺ¡ÒÃà¢éÒáÅжʹÃËÑÊ RLL Ẻ ¢éÍÁÙÅÍÔ¹¾Øµ
¢éÍÁÙÅàÍÒµì¾Øµ
00
010
01
011
10
101
11
110
R=
(0, 2)
2 3
㹷ӹͧà´ÕÂǡѹ ¡ÒÃËÒ»ÃÐÊÔ·¸Ô¼Å¢Í§ÃËÑÊ RLL ÊÒÁÒöËÒä´é µÒÁ¢Ñ ¹µÍ¹µèÍä»¹Õ àÃÔ Áµé¹ ãËéÊÃéÒ§à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´¢Í§ÃËÑÊ RLL Ẻ
(0, 2)
«Ö §¨Ðä´éµÒÁÃÙ»·Õ 8.5 ¨Ò¡¹Ñ ¹ãËéÊÃéÒ§àÁ·ÃÔ¡«ì
¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð·Õ ÊÍ´¤Åéͧ¡Ñº à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ ã¹ÃÙ» ·Õ 8.5 «Ö § ¨Ðä´é ¼ÅÅѾ¸ì µÒÁÊÁ¡Òà (8.20) ¹Ñ ¹¤×Í
1 1 0
D= 1 0 1 1 0 0 ¢Ñ ¹µÍ¹µèÍÁÒ ¤×Í ¡ÒÃËÒ¤èÒ ÅѡɳÐ੾ÒТͧàÁ·ÃÔ¡«ì
D
«Ö § ÊÒÁÒöËÒä´é ¨Ò¡á¡é ÊÁ¡Òà (8.19)
¹Ñ ¹¤×Í
1 1 0
1 0 0
det 1 0 1 − λ 0 1 0 = 0 1 0 0 0 0 1 1−λ 1 0 det 1 −λ 1 = 0 1 0 −λ −λ3 + λ2 + λ + 1 = 0
(8.22)
164
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
â´Â¡ÒÃá¡éÊÁ¡Òà (8.22) àÃÒ¨Ðä´éÇèÒ
λ = 1.8393, −0.4196 + 0.6063i, −0.4196 − 0.6063i ´Ñ§¹Ñ ¹
λD max = 1.8393
¨Ð¶Ù¡¹ÓÁÒãªé㹡ÒäӹdzËÒ¤ÇÒÁ¨Ø
C(d, k)
µÒÁÊÁ¡Òà (8.18) ¹Ñ ¹¤×Í
C(d, k) = log2 {λmax } = log2 {1.8393} = 0.87916 áÅлÃÐÊÔ·¸Ô¼Å¢Í§ÃËÑÊ
η
ËÒä´é¨Ò¡ÊÁ¡Òà (8.10) «Ö §¨Ðä´éÇèÒ
η=
8.7
R 2/3 = = 0.7583 C(d, k) 0.87916
µÑÇÍÂèÒ§ÃËÑÊ RLL à຺µèÒ§æ
ÃËÑÊ RLL ÁÕËÅÒÂẺ¢Ö ¹ÍÂÙè¡Ñº¾ÒÃÒÁÔàµÍÃì
(d, k)
áÅÐÍѵÃÒÃËÑÊ
R
·Õ ãªé ã¹ÂؤàÃÔ Áµé¹¢Í§ÍØ»¡Ã³ì
ÎÒÃì´´ÔÊ¡ìä´Ã¿ì ÃËÑÊ RLL ·Õ ãªé¨ÐÁÕª× ÍÇèÒ ÃËÑÊ FM (frequency modulation) â´ÂÁÕµÒÃÒ§¤é¹ËÒ ÊÓËÃѺ¡ÒÃà¢éÒáÅжʹÃËÑÊ µÒÁÃÙ»·Õ 8.6(a) â´ÂÃËÑÊ FM ¹Õ ¨Ðãªé§Ò¹ÃèÇÁ¡ÑºÇ§¨ÃµÃǨËҨشÊÙ§ÊØ´ (peak detector) áÅÐÁÕÍѵÃÒÃËÑÊ
R = 1/2 «Ö §¨Ð·ÓãËéµéͧÊÙàÊÕÂ¾× ¹·Õ ã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ìä»»ÃÐÁÒ³
50% à¾× Íà¡çº¢éÍÁÙźԵÊèǹà¡Ô¹ µÑÇÍÂèÒ§¡ÒÃà¢éÒÃËÑÊàªè¹ ¶éÒÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ ¤×Í ¢éÍÁÙÅ·Õ ä´é¨Ò¡¡ÒÃà¢éÒÃËÑÊ FM ¤×Í
{110000}
ÅӴѺ
{11 11 01 01 01 01}
Êѧࡵ¨Ð¾ºÇèÒ ÃËÑÊ FM ÂÍÁãËéÅӴѺ¢éÍÁÙÅ·Õ à¢éÒÃËÑÊáÅéÇÁÕºÔµ 1 µÔ´¡Ñ¹ä´é «Ö §¨Ð¡èÍãËéà¡Ô´» ËÒ àÃ× Í§ ISI ´Ñ§¹Ñ ¹ ¨Ö§ÁÕ¡ÒþѲ¹ÒÃËÑÊãËÁè·Õ àÃÕ¡ÇèÒ ÃËÑÊ MFM (modi ed frequency modulation) ËÃ×ͺҧ¤ÃÑ §àÃÕ¡ÇèÒ ÃËÑÊÁÔÅàÅÍÃì (Miller code) «Ö §ÁÕµÒÃÒ§¤é¹ËÒÊÓËÃѺ¡ÒÃà¢éÒáÅжʹÃËÑÊ µÒÁ ÃÙ» ·Õ 8.6(b) â´Â·Õ
x = 0
¶×Í ÇèÒ à» ¹ ÃËÑÊ RLL Ẻ
0.5/0.5515 = 0.9066
¶éÒ ºÔµ ¡è͹˹éÒ ºÔµ
(1, 3)
x
ÁÕ ¤èÒ à» ¹ ºÔµ 1 ¹Í¡¹Ñ ¹
áÅÐÁÕ ÍѵÃÒÃËÑÊ
R = 1/2
{01 01 00 10 10 01 01}
ÃËÑÊ MFM ¹Õ
â´Â¨ÐÁÕ »ÃÐÊÔ·¸Ô¼Å¢Í§ÃËÑÊ
µÑÇÍÂèÒ§¡ÒÃà¢éÒÃËÑÊ àªè¹ ¶éÒÅӴѺ¢éÍÁÙÅÍÔ¹¾Øµ ¤×Í
·Õ ä´é¨Ò¡¡ÒÃà¢éÒÃËÑÊ MFM ¤×Í
x = 1
{1100011}
η =
ÅӴѺ¢éÍÁÙÅ
à» ¹µé¹ ¹Í¡¨Ò¡¹Õ ÃÙ»·Õ 8.6(c) áÅÐ 8.6(d)
áÊ´§µÑÇÍÂèÒ§ÃËÑÊ RLL ẺµèÒ§æ ·Õ ãªéã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì ¨Ò¡ÇÔÇѲ¹Ò¡Òâͧ¡ÒþѲ¹ÒÃËÑÊ RLL
8.7.
µÑÇÍÂèÒ§ÃËÑÊ RLL à຺µèÒ§æ
165
user bits
coded bits
user bits
coded bits
0 1
01 11
0 1
x0 01
(a) FM code
(b) MFM code
user bits
coded bits
user bits
coded bits
00 01 10 11 0000 0001 1000 1001
101 100 001 010 101000 100000 001000 010000
10 11 000 010 011 0010 0011
0100 1000 000100 100100 001000 00100100 00001000
(c) 2/3 (1,7) RLL code
(d) 1/2 (1,7) RLL code
ÃÙ»·Õ 8.6: µÑÇÍÂèÒ§ÃËÑÊ RLL ẺµèÒ§æ ·Õ ãªéã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì
¨Ð¾ºÇèÒ ¾ÒÃÒÁÔàµÍÃì
d
·Õ ãªéã¹ÃËÑÊ RLL ¨Ð¤èÍÂæ Ŵŧ à¾× ÍÅ´¨Ó¹Ç¹ºÔµÊèǹà¡Ô¹ ·ÓãËéÊÒÁÒö
¨Ñ´à¡çº¢éÍÁÙÅ·Õ µéͧ¡ÒÃã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ìä´éÁÒ¡¢Ö ¹ 㹡ÒõѴÊÔ¹ã¨ÇèÒ¨ÐàÍÒÃËÑÊ RLL ã´ÁÒãªé§Ò¹ã¹Ãкº¨Ð¢Ö ¹ÍÂÙè¡Ñº» ¨¨ÑÂËÅÒÂæ ÍÂèÒ§ ´Ñ§¹Õ 1) ¾ÒÃÒÁÔàµÍÃì
2) ÍѵÃÒÃËÑÊ
3) ¤ÇÒÁ¨Ø
(d, k)
R = m/n
C(d, k)
4) »ÃÐÊÔ·¸Ô¼Å¢Í§ÃËÑÊ
η
5) ÍѵÃÒ¤ÇÒÁ˹Òá¹è¹ DR
166
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
«Ö §â´Â·Ñ Çä»áÅéÇ ¨Óà» ¹µéͧ»ÃйջÃйÍÁ» ¨¨ÑÂ·Ñ §ËÁ´ãËéàËÁÒÐÊÁ¡ÑºÊÀÒ¾áÇ´ÅéÍÁ㹡Ò÷ӧҹ ¢Í§Ãкº ÃËÑÊ RLL ·Õ à¤Âãªéã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì àªè¹ ÃËÑÊ RLL Ẻ
4/5 (0, 2),
áÅÐÃËÑÊ RLL Ẻ
8/9 (0, 3)
1/2 (2, 7),
ÃËÑÊ RLL Ẻ
à» ¹µé¹ Êѧࡵ¨Ð¾ºÇèÒ ÃËÑÊ RLL ·Õ ãªé¨ÐÁÕÍѵÃÒÃËÑÊà¢éÒ
ã¡Åé¤èÒ 1 àÃ× ÍÂæ à¾× ÍÅ´¨Ó¹Ç¹ºÔµÊèǹà¡Ô¹ áÅоÒÃÒÁÔàµÍÃì
d
·Õ ãªé¡ç¨ÐÁÕ¤èÒà» ¹¤èÒ 0 ¹Ñ ¹¤×Í ÂÍÁãËé
ÁÕºÔµ 1 µÔ´¡Ñ¹ä´é «Ö §¶Ö§áÁéÇèҨСèÍãËéà¡Ô´» ËÒàÃ× Í§ ISI áµèÃкº¡çÊÒÁÒö¨Ñ´¡ÒáѺ ISI ¹Õ ä´é´éÇ ෤¹Ô¤ PRML µÒÁ·Õ ͸ԺÒÂ㹺··Õ 4
8.8
ÃËÑÊ
(0, G/I)
ÊÓËÃѺªèͧÊÑÒ³ PRML
ÊÓËÃѺªèͧÊÑÒ³ PRML ·Õ ãªé·ÒÃìà¡çµáºº PR4, (·Õ äÁèÁÕÊÑҳú¡Ç¹) ³ àÇÅÒ
k
H(D) = 1 − D2 ,
¢éÍÁÙÅàÍÒµì¾ØµªèͧÊÑÒ³
ÁÕ¤èÒà·èҡѺ¼ÅµèÒ§ÃÐËÇèÒ§¢éÍÁÙÅÍÔ¹¾Øµ 2 µÑÇ ³ àÇÅÒ
k
áÅÐ
k−2
´Ñ§¹Ñ ¹ ªèͧÊÑÒ³¹Õ ¨ÐÁÕ ¤Ø³ÊÁºÑµÔ ¾ÔàÈÉ·Õ ÇèÒ ÅӴѺ ¢éÍÁÙÅ ÂèÍÂàÅ¢¤Õ (odd subsequence) ¨Ðà» ¹ ÍÔÊÃШҡÅӴѺ¢éÍÁÙÅÂèÍÂàÅ¢¤Ùè (even subsequence) ´Ñ§¹Ñ ¹ à¾× Í·Õ ¨ÐŴ˹èǤÇÒÁ¨ÓàÊé¹·Ò§ (path memory) ¢Í§Ç§¨ÃµÃǨËÒÇÕà·ÍÃìºÔ ¨Ó¹Ç¹¢Í§ºÔµ 0 ·Õ àÃÕ§µÔ´µèÍ ¡Ñ¹ ¢Í§áµèÅÐÅӴѺ ¢éÍÁÙÅ ÂèÍ¨РµéͧÁÕä´éäèÁèà¡Ô¹
I
µÑÇ áÅÐà¾× ͪèÇ·ÓãËéÃкºä·ÁÁÔ §ÃԤѿàÇÍÐÃÕÊÒÁÒö·Ó§Ò¹ä´éÍÂèÒ§ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾
áÅШӹǹ¢Í§ºÔµ 0 ·Õ àÃÕ§µÔ´µè͡ѹã¹ÅӴѺ¢éÍÁÙÅÃÇÁ (ÅӴѺ¢éÍÁÙÅÂèÍÂàÅ¢¤Õ ÁÒÃÇÁ¡ÑºÅӴѺ¢éÍÁÙÅ ÂèÍÂàÅ¢¤Ùè) ¨ÐµéͧÁÕä´éäèÁèà¡Ô¹ 0 ã¹ÃËÑÊ
(0, G/I)
G
µÑÇ àËÁ×͹¡Ñº¾ÒÃÒÁÔàµÍÃì
k
ã¹ÃËÑÊ RLL Ẻ
(d, k)
ÊÓËÃѺ¤èÒ
¹Ñ ¹¨ÐËÁÒ¶֧ ÃкºÍ¹ØÒµãËéÅӴѺ¢éÍÁÙÅÃÇÁÊÒÁÒöÁÕºÔµ 1 àÃÕ§µÔ´µè͡ѹ
ä´é àËÁ×͹¡Ñº¾ÒÃÒÁÔàµÍÃì
d
ã¹ÃËÑÊ RLL Ẻ
(d, k)
[9, 59, 60] à¾ÃÒÐÇèÒ Ç§¨ÃµÃǨËÒ PRML
ÁÕ¤ÇÒÁÊÒÁÒö㹨Ѵ¡ÒáѺ ISI ·Õ à¡Ô´¢Ö ¹ä´é ¶éÒ¡Ó˹´ãËé
γ = {γ1 , γ2 , . . . , γn } à» ¹ÅӴѺ¢éÍÁÙÅẺ亹ÒÃÕ·Õ ÁÕ¤ÇÒÁÂÒÇ n ºÔµ à¾ÃÒÐ©Ð¹Ñ ¹
ÅӴѺ¢éÍÁÙÅÂèÍÂàÅ¢¤Õ
γo
áÅÐÅӴѺ¢éÍÁÙÅÂèÍÂàÅ¢¤Ùè
γe
¨Ð¹ÔÂÒÁâ´Â [60]
γ o = {γ1 , γ3 , γ5 , . . . , γ2dn/2e−1 } γ e = {γ2 , γ4 , γ6 , . . . , γ2bn/2c } àÁ× Í
bxc
ÂèÍÂàÅ¢¤Õ
á·¹¨Ó¹Ç¹àµçÁºÇ¡·Õ ÁÒ¡·Õ ÊØ´ ·Õ ÁÕ¤èÒ¹éÍ¡ÇèÒËÃ×Íà·èҡѺ¤èÒ
γo
áÅÐÅӴѺ¢éÍÁÙÅÂèÍÂàÅ¢¤Ùè
γe
¨ÐÁÕ¤ÇÒÁÂÒÇà·èҡѺ
dn/2e
x
´Ñ§¹Ñ ¹ ¨Ðä´éÇèÒ ÅӴѺ¢éÍÁÙÅ
áÅÐ
bn/2c
µÒÁÅӴѺ ÅӴѺ
8.9.
ÊÃØ»·éÒº·
¢éÍÁÙÅ
γ
ÃÇÁ
γ
¨Ð¶Ù¡àÃÕ¡ÇèÒà» ¹ÅӴѺ¢éÍÁÙÅ·Õ ÊÍ´¤Åéͧ¡Ñºà§× ͹䢺ѧ¤Ñº
ÁÕºÔµ 0 àÃÕ§µÔ´µè͡ѹä´éäÁèà¡Ô¹
¡Ñ¹ä´éäÁèà¡Ô¹
I
µÑÇÍÂèÒ§·Õ 8.5 RLL Ẻ
ÇÔ¸Õ·Ó
167
µÑÇ àÁ× Í
G
áÅÐ
¡Ó˹´ãËé
(0, G/I)
I
G
γ
γ = {110100101000110011} G
áÅÐ
·Õ ¡Ó˹´ãËéÁÒ ¨Ðä´éÇèÒ
I
¢Í§
G=3
áÅÐ
áÅÐ
γe
ÁÕºÔµ 0 àÃÕ§µÔ´µèÍ
I=4
¤×Í ÅӴѺ ¢éÍÁÙÅ ·Õ ¶Ù¡ à¢éÒ ÃËÑÊ ´éÇÂÃËÑÊ
γ
γ o = {100110101}
à¾ÃÒÐ©Ð¹Ñ ¹ ¨Ó¹Ç¹ºÔµ 0 ·Õ àÃÕ§µÔ´¡Ñ¹ ÁÒ¡·Õ ÊØ´ ¢Í§ ´Ñ§¹Ñ ¹ ¨Ð¾ºÇèÒ
γo
¡çµèÍàÁ× Í ÅӴѺ¢éÍÁÙÅ
à» ¹àÅ¢¨Ó¹Ç¹àµçÁºÇ¡
¨§ËÒ¤èÒ¾ÒÃÒÁÔàµÍÃì
¨Ò¡ÅӴѺ¢éÍÁÙÅ
µÑÇ áÅÐÅӴѺ¢éÍÁÙÅÂèÍÂ
(0, G/I)
γ, γo,
¹Ñ ¹¤×Í ÅӴѺ¢éÍÁÙÅ
γ
áÅÐ
γe
áÅÐ
γ e = {110000101}
¤×Í 3, 2, áÅÐ 4 µÒÁÅӴѺ
¶Ù¡à¢éÒÃËÑÊ´éÇÂÃËÑÊ RLL Ẻ
¨Ò¡µÑÇÍÂèÒ§·Õ 8.5 ¨ÐÊѧࡵàËç¹ä´éÇèÒ ÅӴѺ ¢éÍÁÙÅ ·Õ ¶Ù¡ à¢éÒ ÃËÑÊ ´éÇÂÃËÑÊ RLL Ẻ ·Õ ¶Ù¡µéͧ¹Ñ ¹ ¤èÒ ¾ÒÃÒÁÔàµÍÃì
G ≤ 2I
ÊÑÒ³¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ì àªè¹ ÃËÑÊ
8.9
àÊÁÍ µÑÇÍÂèÒ§ÃËÑÊ
8/9 (0, 4/4)
(0, G/I)
áÅÐÃËÑÊ
(0, 3/4)
(0, G/I)
·Õ ãªé ã¹Ãкº¡ÒûÃÐÁÇżÅ
16/17 (0, 6/6)
à» ¹µé¹
ÊÃØ»·éÒº·
â´Â·Ñ Çä» ÃËÑÊ RLL ÁÕ ËÅÒÂÃٻẺ 㹺·¹Õ ä´é ͸ԺÒ¶֧ ËÅÑ¡¡Ò÷ӧҹáÅÐ¢Ñ ¹µÍ¹¡ÒÃÍ͡Ẻ ÃËÑÊ RLL (run length limited) µÒÁà§× ͹䢺ѧ¤Ñº
(d, k)
â´Â·Õ ¾ÒÃÒÁÔàµÍÃì
d
¨Ðà» ¹ µÑÇ¡Ó˹´
¨Ó¹Ç¹·Õ ¹éÍÂ·Õ ÊØ´ ¢Í§ºÔµ 0 ·Õ ÍÂÙè ÃÐËÇèÒ§ºÔµ 1 (µÒÁÃٻẺ NRZI) áÅоÒÃÒÁÔàµÍÃì
k
¨Ðà» ¹
µÑÇ¡Ó˹´ÁÒ¡·Õ ÊØ´¢Í§ºÔµ 0 ·Õ ÍÂÙèÃÐËÇèÒ§ºÔµ 1 à¹× ͧ¨Ò¡ ºÔµ 1 á·¹¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð¢Í§¡ÃÐáÊ ä¿¿ Ò à¢Õ¹ ´Ñ§¹Ñ ¹ ¾ÒÃÒÁÔàµÍÃì
d
¨Ö§ ªèÇÂÅ´¼Å¡Ãзº·Õ à¡Ô´ ¨Ò¡ ISI ã¹¢³Ð·Õ ¾ÒÃÒÁÔàµÍÃì
k
¨Ð
ÃѺ»ÃСѹÇèÒ ÅӴѺ¢éÍÁÙÅ·Õ à¢Õ¹ŧã¹Ê× ÍºÑ¹·Ö¡¨ÐÁÕ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ðà» ¹ÃÐÂÐæ à¾× ͪèÇ·ÓãËéÃкº ä·ÁÁÔ §ÃԤѿàÇÍÐÃÕÊÒÁÒö·Ó§Ò¹ä´éÍÂèÒ§ÁÕ»ÃÐÊÔ·¸ÔÀÒ¾ ÍÂèÒ§äáçµÒÁ ã¹» ¨¨ØºÑ¹¹Õ ÃËÑÊ RLL ·Õ ¹ÔÂÁ ãªéÁÑ¡¨ÐÍÂÙèã¹ÃÙ»¢Í§ÃËÑÊ ´ÔÊ¡ìä´Ã¿ì
(0, G/I)
«Ö §à» ¹ÃËÑÊ·Õ ¶Ù¡Í͡ẺÁÒãËéãªé§Ò¹¡ÑºÃкº PRML ¢Í§ÎÒÃì´
168
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
㹡ÒÃàÅ×Í¡ÃËÑÊ RLL ÁÒãªé§Ò¹¨Ðµéͧ¤Ó¹Ö§¶Ö§» ¨¨ÑµèÒ§æ ä´éá¡è ¾ÒÃÒÁÔàµÍÃì
R,
¤ÇÒÁ¨Ø
C(d, k),
»ÃÐÊÔ·¸Ô¼Å¢Í§ÃËÑÊ
η,
(d, k),
ÍѵÃÒÃËÑÊ
áÅÐ ÍѵÃÒ¤ÇÒÁ˹Òá¹è¹ DR à» ¹µé¹ à¾× ÍãËéä´éÃËÑÊ
RLL ·Õ ´ÕÊØ´ÊÓËÃѺ§Ò¹»ÃÐÂØ¡µì¹Ñ ¹æ ¢éÍÊѧࡵ·Õ ¾º¢Í§ÃËÑÊ RLL ·Õ ¹ÓÁÒãªé§Ò¹ã¹ÎÒÃì´´ÔÊ¡ìä´Ã¿ì ã¹» ¨¨ØºÑ¹ ¤×Í ¾ÒÃÒÁÔàµÍÃì
d
¨Ð¤èÍÂæ ŴŧÁÒà» ¹¤èÒ 0 áÅÐÍѵÃÒÃËÑÊ
R
¢Í§ÃËÑÊ RLL ·Õ ãªé ¡çÁÕ
¤èÒà¢éÒã¡Åé¤èÒ 1 ÁÒ¡¢Ö ¹àÃ× ÍÂæ ·Ñ §¹Õ à¾× Íà» ¹¡ÒÃÅ´¨Ó¹Ç¹ºÔµÊèǹà¡Ô¹·Õ à¡Ô´¢Ö ¹ã¹Ãкº ·ÓãËéÊÒÁÒö ¨Ñ´à¡çº¢éÍÁÙÅ¢èÒÇÊÒâͧ¼Ùéãªéä´éÁÒ¡¢Ö ¹
8.10
à຺½ ¡ËÑ´·éÒº·
1. ¨§¤Ó¹Ç³ËҨӹǹÅӴѺ¢éÍÁÙÅ·Ñ §ËÁ´·Õ ÁÕ¤ÇÒÁÂÒÇ·Ñ §ÊÔ ¹
(d, k)
L ºÔµ ·Õ ÊÍ´¤Åéͧ¡Ñºà§× ͹䢺ѧ¤Ñº
àÁ× Í
1.1)
d = 0, k = 2,
áÅÐ
L=5
1.2)
d = 1, k = 3,
áÅÐ
L=6
1.3)
d = 1, k = 7,
áÅÐ
L = 10
1.4)
d = 2, k = 7,
áÅÐ
L = 10
1.5)
d = 2, k = ∞,
áÅÐ
L = 10
2. ¨§áÊ´§á¼¹ÀÒ¾à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´ áÅФӹdzËÒàÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð ¢Í§ÃËÑÊ RLL µÒÁà§× ͹䢺ѧ¤Ñº¢Í§¾ÒÃÒÁÔàµÍÃì
2.1)
d=1
áÅÐ
k=5
2.2)
d=1
áÅÐ
k=7
2.3)
d=2
áÅÐ
k=5
2.4)
d=2
áÅÐ
k=7
(d, k)
3. ¨Ò¡â¨·Âìã¹¢éÍ·Õ 2 ¨§¤Ó¹Ç³ËÒ¤ÇÒÁ¨Ø ¼ÅÅѾ¸ì·Õ äè´é¡Ñº
C(d, k)
àÁ× Í
C(d, k)
â´ÂãªéÊÁ¡Òà (8.12) ¾ÃéÍÁ·Ñ §à»ÃÕºà·Õº
·Õ ¤Ó¹Ç³ä´é¨Ò¡ÊÁ¡Òà (8.18)
8.10.
à຺½ ¡ËÑ´·éÒº·
169
4. ¨§ËҨӹǹÅӴѺ ¢éÍÁÙÅ ·Õ à» ¹ ä»ä´é ·Ñ §ËÁ´·Õ ÁÕ ¤ÇÒÁÂÒÇ ÊÔ ¹ÊØ´·Õ ʶҹÐ
Sj
d = 0, k = 2,
áÅÐ
L=5
4.2)
d = 1, k = 3,
áÅÐ
L=5
4.3)
d = 1, k = 7,
áÅÐ
L = 10
4.4)
d = 2, k = 7,
áÅÐ
L = 10
5. ¡Ó˹´ãËéÅӴѺ¢éÍÁÙÅ·Õ ¶Ù¡à¢éÒÃËÑÊ´éÇÂÃËÑÊ RLL Ẻ ¢Í§
γ
ºÔµ ·Õ ÍÍ¡¨Ò¡Ê¶Ò¹Ð
¢Í§ÃËÑÊ RLL µÒÁà§× ͹䢺ѧ¤Ñº¢Í§¾ÒÃÒÁÔàµÍÃì
4.1)
I
L
(0, G/I)
(d, k)
Si
áÅéÇ ä»
àÁ× Í
¨§ËÒ¤èÒ¾ÒÃÒÁÔàµÍÃì
G
áÅÐ
´Ñ§µèÍ仹Õ
5.1)
γ = {10101000101001100101}
5.2)
γ = {10010100001101001100111101}
5.3)
γ = {11100011001010001001100101}
6. ¨§ÊÃéÒ§µÒÃÒ§¤é¹ËÒÊÓËÃѺ ¡ÒÃà¢éÒ áÅжʹÃËÑÊ RLL Ẻ ¼èÒ¹¡ÒÃà¢éÒÃËÑÊáÅéǨÐÁÕ¤ÇÒÁÂÒÇà·èҡѺ
L=2
L=3
L=4
(1, 2)
â´Â·Õ ¢éÍÁÙÅ áµèÅФÃÑ § ·Õ
¾ÃéÍÁ·Ñ §ËÒ»ÃÐÊÔ·¸Ô¼Å¢Í§ÃËÑÊ
8. ¨§ÊÃéÒ§µÒÃÒ§¤é¹ËÒÊÓËÃѺ ¡ÒÃà¢éÒ áÅжʹÃËÑÊ RLL Ẻ ¼èÒ¹¡ÒÃà¢éÒÃËÑÊáÅéǨÐÁÕ¤ÇÒÁÂÒÇà·èҡѺ
â´Â·Õ ¢éÍÁÙÅ áµèÅФÃÑ § ·Õ
¾ÃéÍÁ·Ñ §ËÒ»ÃÐÊÔ·¸Ô¼Å¢Í§ÃËÑÊ
7. ¨§ÊÃéÒ§µÒÃÒ§¤é¹ËÒÊÓËÃѺ ¡ÒÃà¢éÒ áÅжʹÃËÑÊ RLL Ẻ ¼èÒ¹¡ÒÃà¢éÒÃËÑÊáÅéǨÐÁÕ¤ÇÒÁÂÒÇà·èҡѺ
(0, 1)
(1, 3)
â´Â·Õ ¢éÍÁÙÅ áµèÅФÃÑ § ·Õ
¾ÃéÍÁ·Ñ §ËÒ»ÃÐÊÔ·¸Ô¼Å¢Í§ÃËÑÊ
170
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ÀÒ¤¼¹Ç¡ ¡
µÒÃÒ§¿ §¡ìªÑ¹
¿ §¡ìªÑ¹
Q(x)
Q
à» ¹¿ §¡ìªÑ¹·Õ ÊÒÁÒö¨Ñ´ãËéÍÂÙèã¹ÃÙ»¢Í§¿ §¡ìªÑ¹¡ÒÃᨡᨧÊÐÊÁ¢Í§µÑÇá»ÃÊØèÁẺ
à¡ÒÊìà«Õ¹ä´é«Ö §à» ¹·Õ ¹ÔÂÁãªé§Ò¹·Ò§´éҹʶԵÔÈÒʵÃìáÅдéÒ¹ÇÔÈÇ¡ÃÃÁÈÒʵÃì â´Â¿ §¡ìªÑ¹ ¹ÔÂÒÁ´Ñ§¹Õ
1 Q(x) = √ 2π
Z
∞
x
Q(x)
½ 2¾ y dy exp − 2
(¡.1)
«Ö §à» ¹¡ÒÃËÒ¤èÒ»ÃԾѹ¸ìÊèǹËÒ§¢Í§¿ §¡ìªÑ¹¤ÇÒÁ˹Òá¹è¹¤ÇÒÁ¹èÒ¨Ðà» ¹áººà¡ÒÊìà«Õ¹àÁ× Í ¤×Í ¿ §¡ìªÑ¹àÅ¢ªÕ ¡ÓÅѧ (exponential function) â´Â·Ñ Çä» ¤èҢͧ¿ §¡ìªÑ¹ ÊÒÁÒöËÒä´é¨Ò¡µÒÃÒ§¤é¹ËÒ (look up table) áµèã¹¡Ã³Õ·Õ ¤èÒä´é´Ñ§¹Õ
xÀ3
½ 2¾ 1 x Q(x) ≈ √ exp − 2 x 2π
µÒÃÒ§µèÍä»¹Õ ¨ÐáÊ´§¤èҢͧ¿ §¡ìªÑ¹
Q(x)
ÊÓËÃѺ
0 ≤ x ≤ 3.59
171
¨Ð
Q(x)
¿ §¡ìªÑ¹
ÊÓËÃѺ¤èÒ
Q(x)
exp{·}
x
µèÒ§æ
ÊÒÁÒö»ÃÐÁÒ³
(¡.2)
172
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
x
Q(x)
0
0.50000
0.36
x
Q(x) 0.35942
0.72
x
Q(x) 0.23576
1.08
x
Q(x) 0.14007
1.44
x
Q(x) 0.0749340
0.01
0.49601
0.37
0.35569
0.73
0.23270
1.09
0.13786
1.45
0.0735290
0.02
0.49202
0.38
0.35197
0.74
0.22965
1.10
0.13567
1.46
0.0721450
0.03
0.48803
0.39
0.34827
0.75
0.22663
1.11
0.133500
1.47
0.070781
0.04
0.48405
0.40
0.34458
0.76
0.22363
1.12
0.131360
1.48
0.069437
0.05
0.48006
0.41
0.34090
0.77
0.22065
1.13
0.129240
1.49
0.068112
0.06
0.47608
0.42
0.33724
0.78
0.21770
1.14
0.127140
1.50
0.066807
0.07
0.47210
0.43
0.33360
0.79
0.21476
1.15
0.125070
1.51
0.065522
0.08
0.46812
0.44
0.32997
0.80
0.21186
1.16
0.123020
1.52
0.064255
0.09
0.46414
0.45
0.32636
0.81
0.20897
1.17
0.121000
1.53
0.063008
0.10
0.46017
0.46
0.32276
0.82
0.20611
1.18
0.119000
1.54
0.061780
0.11
0.45620
0.47
0.31918
0.83
0.20327
1.19
0.117020
1.55
0.060571
0.12
0.45224
0.48
0.31561
0.84
0.20045
1.20
0.115070
1.56
0.059380
0.13
0.44828
0.49
0.31207
0.85
0.19766
1.21
0.113140
1.57
0.058208
0.14
0.44433
0.50
0.30854
0.86
0.19489
1.22
0.111230
1.58
0.057053
0.15
0.44038
0.51
0.30503
0.87
0.19215
1.23
0.109350
1.59
0.055917
0.16
0.43644
0.52
0.30153
0.88
0.18943
1.24
0.107490
1.60
0.054799
0.17
0.43251
0.53
0.29806
0.89
0.18673
1.25
0.105650
1.61
0.053699
0.18
0.42858
0.54
0.29460
0.90
0.18406
1.26
0.103830
1.62
0.052616
0.19
0.42465
0.55
0.29116
0.91
0.18141
1.27
0.102040
1.63
0.051551
0.20
0.42074
0.56
0.28774
0.92
0.17879
1.28
0.100270
1.64
0.050503
0.21
0.41683
0.57
0.28434
0.93
0.17619
1.29
0.098525
1.65
0.049471
0.22
0.41294
0.58
0.28096
0.94
0.17361
1.30
0.096800
1.66
0.048457
0.23
0.40905
0.59
0.27760
0.95
0.17106
1.31
0.095098
1.67
0.047460
0.24
0.40517
0.60
0.27425
0.96
0.16853
1.32
0.093418
1.68
0.046479
0.25
0.40129
0.61
0.27093
0.97
0.16602
1.33
0.091759
1.69
0.045514
0.26
0.39743
0.62
0.26763
0.98
0.16354
1.34
0.090123
1.70
0.044565
0.27
0.39358
0.63
0.26435
0.99
0.16109
1.35
0.088508
1.71
0.043633
0.28
0.38974
0.64
0.26109
1.00
0.15866
1.36
0.086915
1.72
0.042716
0.29
0.38591
0.65
0.25785
1.01
0.15625
1.37
0.085343
1.73
0.041815
0.30
0.38209
0.66
0.25463
1.02
0.15386
1.38
0.083793
1.74
0.040930
0.31
0.37828
0.67
0.25143
1.03
0.15151
1.39
0.082264
1.75
0.040059
0.32
0.37448
0.68
0.24825
1.04
0.14917
1.40
0.080757
1.76
0.039204
0.33
0.37070
0.69
0.24510
1.05
0.14686
1.41
0.079270
1.77
0.038364
0.34
0.36693
0.70
0.24196
1.06
0.14457
1.42
0.077804
1.78
0.037538
0.35
0.36317
0.71
0.23885
1.07
0.14231
1.43
0.076359
1.79
0.036727
173
x
Q(x)
x
Q(x)
x
Q(x)
x
Q(x)
x
Q(x)
1.80
0.035930
2.16
0.0153860
2.52
0.0058677
2.88
0.00198840
3.24
0.00059765
1.81
0.035148
2.17
0.0150030
2.53
0.0057031
2.89
0.00192620
3.25
0.00057703
1.82
0.034380
2.18
0.0146290
2.54
0.0055426
2.90
0.00186580
3.26
0.00055706
1.83
0.033625
2.19
0.0142620
2.55
0.0053861
2.91
0.00180710
3.27
0.00053774
1.84
0.032884
2.20
0.0139030
2.56
0.0052336
2.92
0.00175020
3.28
0.00051904
1.85
0.032157
2.21
0.0135530
2.57
0.0050849
2.93
0.00169480
3.29
0.00050094
1.86
0.031443
2.22
0.0132090
2.58
0.0049400
2.94
0.00164110
3.30
0.00048342
1.87
0.030742
2.23
0.0128740
2.59
0.0047988
2.95
0.00158890
3.31
0.00046648
1.88
0.030054
2.24
0.0125450
2.60
0.0046612
2.96
0.00153820
3.32
0.00045009
1.89
0.029379
2.25
0.0122240
2.61
0.0045271
2.97
0.00148900
3.33
0.00043423
1.90
0.028717
2.26
0.0119110
2.62
0.0043965
2.98
0.00144120
3.34
0.00041889
1.91
0.028067
2.27
0.0116040
2.63
0.0042692
2.99
0.00139490
3.35
0.00040406
1.92
0.027429
2.28
0.0113040
2.64
0.0041453
3.00
0.00134990
3.36
0.00038971
1.93
0.026803
2.29
0.0110110
2.65
0.0040246
3.01
0.00130620
3.37
0.00037584
1.94
0.026190
2.30
0.0107240
2.66
0.0039070
3.02
0.00126390
3.38
0.00036243
1.95
0.025588
2.31
0.0104440
2.67
0.0037926
3.03
0.00122280
3.39
0.00034946
1.96
0.024998
2.32
0.0101700
2.68
0.0036811
3.04
0.00118290
3.40
0.00033693
1.97
0.024419
2.33
0.0099031
2.69
0.0035726
3.05
0.00114420
3.41
0.00032481
1.98
0.023852
2.34
0.0096419
2.70
0.0034670
3.06
0.00110670
3.42
0.00031311
1.99
0.023295
2.35
0.0093867
2.71
0.0033642
3.07
0.00107030
3.43
0.00030179
2.00
0.022750
2.36
0.0091375
2.72
0.0032641
3.08
0.00103500
3.44
0.00029086
2.01
0.022216
2.37
0.0088940
2.73
0.0031667
3.09
0.00100080
3.45
0.00028029
2.02
0.021692
2.38
0.0086563
2.74
0.0030720
3.10
0.00096760
3.46
0.00027009
2.03
0.021178
2.39
0.0084242
2.75
0.0029798
3.11
0.00093544
3.47
0.00026023
2.04
0.020675
2.40
0.0081975
2.76
0.0028901
3.12
0.00090426
3.48
0.00025071
2.05
0.020182
2.41
0.0079763
2.77
0.0028028
3.13
0.00087403
3.49
0.00024151
2.06
0.019699
2.42
0.0077603
2.78
0.0027179
3.14
0.00084474
3.50
0.00023263
2.07
0.019226
2.43
0.0075494
2.79
0.0026354
3.15
0.00081635
3.51
0.00022405
2.08
0.018763
2.44
0.0073436
2.80
0.0025551
3.16
0.00078885
3.52
0.00021577
2.09
0.018309
2.45
0.0071428
2.81
0.0024771
3.17
0.00076219
3.53
0.00020778
2.10
0.017864
2.46
0.0069469
2.82
0.0024012
3.18
0.00073638
3.54
0.00020006
2.11
0.017429
2.47
0.0067557
2.83
0.0023274
3.19
0.00071136
3.55
0.00019262
2.12
0.017003
2.48
0.0065691
2.84
0.0022557
3.20
0.00068714
3.56
0.00018543
2.13
0.016586
2.49
0.0063872
2.85
0.0021860
3.21
0.00066367
3.57
0.00017849
2.14
0.016177
2.50
0.0062097
2.86
0.0021182
3.22
0.00064095
3.58
0.00017180
2.15
0.015778
2.51
0.0060366
2.87
0.0020524
3.23
0.00061895
3.59
0.00016534
174
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ÀÒ¤¼¹Ç¡ ¢
Êٵä³ÔµÈÒʵÃì·Õ ÊÓ¤Ñ
ÀÒ¤¼¹Ç¡¹Õ ¨ÐáÊ´§Êٵä³ÔµÈÒʵÃì ·Õ ãªéºèÍÂÊÓËÃѺ ¡ÒÃÇÔà¤ÃÒÐËì Ãкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ ÎÒÃì´´ÔÊ¡ìä´Ã¿ì¢Í§Ë¹Ñ§Ê×ÍàÅèÁ¹Õ
¢.1
µÃÕ⡳ÁÔµÔ (Trigonometric)
sin(−α) = − sin(α) cos(−α) = cos(α) sin(α) = cos(α − π/2) sin2 (α) + cos2 (α) = 1 sin(α ± β) = sin(α) cos(β) ± cos(α) sin(β) cos(α ± β) = cos(α) cos(β) ∓ sin(α) sin(β) sin(α) sin(β) =
1 2
cos(α − β) − 12 cos(α + β)
sin(α) cos(β) =
1 2
sin(α + β) + 12 sin(α − β)
cos(α) cos(β) =
1 2
cos(α − β) + 21 cos(α + β)
cos(α) sin(β) =
1 2
sin(α + β) − 12 sin(α − β) 175
176
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
sin(2α) = 2 sin(α) cos(α) cos(2α) = cos2 (α) − sin2 (α) = 1 − 2 sin2 (α) = 2 cos2 (α) − 1 sin2 (α) = 21 {1 − cos(2α)} cos2 (α) = 21 {1 + cos(2α)} ejα = cos(α) + j sin(α) sin(α) = (ejα − e−jα )/(2j) cos(α) = (ejα + e−jα )/2
¢.2
R R R R R R R R R R R R R
»ÃԾѹ¸ìäÁè¨Ó¡Ñ´à¢µ (Inde nite Integral)
u dv = uv −
R
v du
àÁ× Í
xn dx = xn+1 /(n + 1)
u
áÅÐ
àÁ× Í
v
à» ¹¿ §¡ìªÑ¹¢Í§
n 6= −1
x−1 dx = ln(x) eax dx = eax /a ln(x) dx = x ln(x) − x xeax dx = eax (ax − 1)/a2 x2 eax dx = eax (a2 x2 − 2ax + 2)/a3 sin(ax) dx = −(1/a) cos(ax) cos(ax) dx = (1/a) sin(ax) sin2 (ax) dx = x/2 − sin(2ax)/4a x sin(ax) dx = (1/a2 ){sin(ax) − ax cos(ax)} cos2 (ax) dx = x/2 + sin(2ax)/4a x cos(ax) dx = (1/a2 ){cos(ax) + ax sin(ax)}
x
ÀÒ¤¼¹Ç¡ ¤
¡ÒÃËÒ͹ؾѹ¸ì¢Í§àÇ¡àµÍÃìààÅÐàÁ·ÃÔ¡«ì
ÀÒ¤¼¹Ç¡¹Õ ¨ÐÊÃØ»ÊٵáÒÃËÒ͹ؾѹ¸ì (di erentiation) ¢Í§àÇ¡àµÍÃìáÅÐàÁ·ÃÔ¡«ì«Ö §¨Ðà» ¹»ÃÐâª¹ì µèÍ¡ÒÃÇÔà¤ÃÒÐËìÃкº¡ÒûÃÐÁÇżÅÊÑÒ³¢Í§ÎÒÃì´´ÔÊ¡ìä´Ã¿ìã¹Ë¹Ñ§Ê×ÍàÅèÁ¹Õ ¡Ó˹´ãËé
k
u
áÅÐ
x
à» ¹àÇ¡àµÍÃìá¹ÇµÑ § (column vector) ¢¹Ò´
á¶Ç áÅÐ 1 á¹ÇµÑ §) áÅÐãËé
A
k×1
à» ¹àÁ·ÃÔ¡«ì¨ÑµØÃÑÊ (square matrix) ¢¹Ò´
(¹Ñ ¹¤×ÍÁըӹǹ
k×k
´Ñ§¹Ñ ¹
¨Ò¡¡ÒÃËÒ͹ؾѹ¸ì¢Í§àÇ¡àµÍÃìààÅÐàÁ·ÃÔ¡«ì¨Ðä´é¤ÇÒÁÊÑÁ¾Ñ¹¸ì´Ñ§¹Õ
xT u = uT x ¡ ¢ ∂ xT u =x ∂u ∂ (Au) = AT ∂u ∂ (Au) =A ∂uT áÅÐ
¶éÒàÁ·ÃÔ¡«ì
(¤.1)
(¤.2)
(¤.3)
(¤.4)
¡ ¢ ¡ ¢ ∂ uT Au = A + AT u ∂u A
à» ¹àÁ·ÃÔ¡«ì·Õ ÊÁÁҵà (symmetric) ¹Ñ ¹¤×Í
A = AT
¡ ¢ ∂ uT Au = 2Au ∂u 177
(¤.5) ¨Ðä´éÇèÒ
(¤.6)
178
áÅÐ
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¡ ¢ ∂ 2 uT Au = 2A ∂u ∂uT
(¤.7)
ÀÒ¤¼¹Ç¡ §
¤ÓÈѾ·ìà·¤¹Ô¤
¡Ãкǹ¡ÒÃà¢Õ¹
write process minimization process
¡Ãкǹ¡Ò÷ÓãËéÁÕ¤èÒ¹éÍÂÊØ´
¡Ãкǹ¡ÒÃ㹡Ò÷ӹÒÂÊÑҳú¡Ç¹
noise prediction process
¡Ãкǹ¡ÒÃ㹡Ò÷ÓãËéÊÑҳú¡Ç¹à» ¹ÊÕ¢ÒÇ ¡Ãкǹ¡ÒÃÍèÒ¹
read process write current
¡ÃÐáÊä¿¿ Òà¢Õ¹ ¡ÅØèÁ¢éÍÁÙÅ
data packet
¡ÒáŠÓÃËÑʾÑÅÊì
PCM (pulse code modulation) modulation
¡ÒáŠÓÊÑÒ³ (¡ÒÃÁÍ´Ùàŵ) ¡ÒâÂÒÂÊÑҳú¡Ç¹
noise enhancement
synchronization
¡ÒÃà¢éҨѧËÇÐ
perfect synchronization
¡ÒÃà¢éҨѧËÇÐÍÂèÒ§ÊÁºÙÃ³ì ¡ÒäҴËÁÒÂ, ¤èÒ¤Ò´ËÁÒ ¡ÒèѴà¡çº¢éÍÁÙÅ
sampling
¡ÒÃà´Ô¹áººÊØèÁ
random walk
acquisition
¡ÒõÃǨËÒÅӴѺ ¡ÒõԴµÒÁ
digital data storage
projection
¡ÒêѡµÑÇÍÂèÒ§
¡ÒÃä´éÁÒ
expectation
data storage
¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅ ¡ÒéÒÂ
noise whitening process
sequence detection
tracking
179
180
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¡Ò÷ӹÒÂÊÑҳú¡Ç¹ ¡ÒÃá·Ã¡ÊÍ´
noise prediction
interference
¡ÒÃá·Ã¡ÊÍ´ÃÐËÇèÒ§ÊÑÅѡɳì
ISI (intersymbol interference)
¡Òúǡ ¡ÒÃà»ÃÕºà·Õº ¡ÒÃàÅ×Í¡ ¡Òúѹ·Ö¡
ACS (add compare select)
recording
¡Òúѹ·Ö¡áººá¹ÇµÑ §
perpendicular recording
¡Òúѹ·Ö¡áººá¹Ç¹Í¹
longitudinal recording binary recording
¡Òúѹ·Ö¡áººäº¹ÒÃÕ ¡Òúѹ·Ö¡ÃкºáÁèàËÅç¡
magnetic recording
¡ÒûÃÐÁÇżÅẺà¾Íà«ÍÃìäÇàÇÍÃì
¡ÒûÃѺ¤èÒ·Ò§àÇÅÒ ¡Òû ͹¡ÅѺ
timing adjustment
feedback
¡Òû ͹¡ÅѺ¤èҵѴÊԹ㨠¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
decision feedback
transition
¡ÒÃà»ÅÕ Â¹Ê¶Ò¹ÐàÍ¡à·È
Z
¡ÒÃá»Å§«Õ
isolated transition
transform
¡ÒÃá»Å§¿ÙàÃÕÂÃì
Fourier transform
¡ÒÃá»Å§¿ÙàÃÕÂÃì·Õ µèÍà¹× ͧ·Ò§àÇÅÒ ¡ÒÃá¾Ãè¡ÃШÒ¢ͧ¢éͼԴ¾ÅÒ´ ¡ÒÃú¡Ç¹
PSP (per survivor processing)
digital signal processing
¡ÒûÃÐÁÇżÅÊÑÒ³´Ô¨Ô·ÑÅ
continuous time Fourier transform error propagation
disturbance
¡ÒÃźÅéÒ§ÊÀÒ¾áÁèàËÅç¡
demagnetization
¡ÒÃàÅ× Í¹µÓá˹觢ͧ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð
transition shift
¡ÒÃàÅ× Í¹µÓá˹觢ͧ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹ÐẺÊØèÁ ¡ÒÃ˹èǧàÇÅÒ¡ÒõѴÊԹ㨠¡ÒÃËÒ͹ؾѹ¸ì ¡ÓÅѧ
decision delay
di erentiation
power
à¡ÒÊìà«Õ¹ ¢¹Ò´
random transition shift
Gaussian magnitude
¢éͼԴ¾ÅÒ´
error prediction error
¢éͼԴ¾ÅÒ´¡Ò÷ӹÒ ¢éͼԴ¾ÅÒ´¡ÓÅѧÊͧà©ÅÕ Â
MSE (mean squared error)
¢éͼԴ¾ÅÒ´¡ÓÅѧÊͧà©ÅÕ Â·Õ ¹éÍÂÊØ´ ¢éͼԴ¾ÅÒ´·Ò§¤ÇÒÁ¶Õ ¢éͼԴ¾ÅÒ´·Ò§à¿Ê
MMSE (minimum mean squared error)
frequency error phase error
181
¢éͼԴ¾ÅÒ´·Ò§àÇÅÒ
timing error
¢éͼԴ¾ÅÒ´·Ò§àÇÅÒ·Õ ËŧàËÅ×ÍÍÂÙè ¢éÍÁÙÅ亹ÒÃÕ
binary data
¢éÍÁÙÅÇÔÂص (á«Áà» Å) ¤ÇèÐà» ¹ÁÒ¡ÊØ´ ¤ÇÒÁ¨Ø
residual timing error
discrete data ML (maximum likelihood)
capacity
¤ÇÒÁ¨ØªèͧÊÑÒ³
channel capacity
complexity
¤ÇÒÁ«Ñº«é͹
¤ÇÒÁ¶Õ ¡ÒêѡµÑÇÍÂèÒ§ ¤ÇÒÁ¶Õ µÑ´
sampling rate
cut o frequency normalized frequency
¤ÇÒÁ¶Õ Ẻ¹ÍÃìÁÍÅäÅ«ì ¤ÇÒÁ¶Õ 乤ÇÔµÊì
Nyquist frequency
¤ÇÒÁ¹èÒ¨Ðà» ¹
probability
¤ÇÒÁ¹èÒ¨Ðà» ¹¢Í§¢éͼԴ¾ÅÒ´
probability of error
¤ÇÒÁ¹èÒ¨Ðà» ¹¢Í§¢éͼԴ¾ÅÒ´ÅӴѺ ¤ÇÒÁá»Ã»Ãǹ
probability of sequence error
variance
¤ÇÒÁäÁèà» ¹àªÔ§àÊé¹
nonlinearity
¤ÇÒÁ˹Òá¹è¹¡ÒúÃèØ
packing density
¤ÇÒÁ˹Òá¹è¹¢Í§¡Òúѹ·Ö¡áºº¹ÍÃìÁÍÅäÅ«ì ¤ÇÒÁ˹Òá¹è¹àªÔ§¾× ¹·Õ
areal density
¤ÇÒÁ˹Òá¹è¹Ê໡µÃÑÁ¡ÓÅѧ ¤Í¹âÇÅ٪ѹ
power spectral density
convolution
¤èÒ¡ÓÅѧÊͧà©ÅÕ Â ¤èÒà©ÅÕ Â
ND (normalized recording density)
mean square
mean
¤èҵѴÊÔ¹ã¨
decision
¤èҵѴÊԹ㨢³ÐË¹Ö §áººá¢ç§ ¤Òº (àÇÅÒ)
instantaneous hard decision
period
¤Òº¡ÒêѡµÑÇÍÂèÒ§ ¤ÒºàÇÅҢͧºÔµ ¤èÒàÁµÃÔ¡ÊÒ¢Ò
sampling period bit period branch metric
¤èÒàÁµÃÔ¡àÊé¹·Ò§
path metric
¤èÒÅѡɳÐ੾ÒÐ
eigenvalue
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´
FSM ( nite state machine)
à¤Ã× Í§ËÁÒÂà¢éҨѧËÇÐ §Ò¹»ÃÐÂØ¡µì (á;ÅÔपѹ)
sync mark application
182
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
à§× ͹䢺ѧ¤Ñº
constraint
à§× ͹䢺ѧ¤ÑºáººâÁ¹Ô¡ ¨Ò¹ (¨Ò¹ºÑ¹·Ö¡)
disk
¨Ò¹ºÑ¹·Ö¡áÁèàËÅç¡ ¨ÔµàµÍÃì
monic constraint
magnetic disk
jitter
¨ÔµàµÍÃì·Ò§àÇÅÒ
timing jitter
node
¨Ø´µèÍ ¨Ø´ÊÁ´ØÅ
equilibrium point
¨Ø´àʶÕÂÃÀÒ¾
stable point
ªèͧÊÑÒ³
channel
ªèͧÊÑÒ³ÇÔÂص
discrete channel
ªèͧÊÑÒ³Ê× ÍÊÒÃ
communication channel
ªèͧÊÑÒ³ÍèÒ¹
read channel
ªÔ»ªèͧÊÑÒ³ÍèÒ¹ àªÔ§àÊé¹
read channel chip
linear
«é͹àËÅ× ÍÁ
overlap superparamagnetic
«Ùà»ÍÃì¾ÒÃÒáÁ¡à¹µÔ¡ à«¡àµÍÃì
sector
á«Áà» Å, µÑÇÍÂèÒ§ ä«à¤ÔÅÊÅÔ» ´Ô¨Ô·ÑÅ
sample
cycle slip digital
´Õà·ÍÃìÁÔá¹¹µì à´«ÔàºÅ
determinant
dB (decibel)
â´àÁ¹
domain
â´àÁ¹
D
D
â´àÁ¹
Z
Z
â´àÁ¹¤ÇÒÁ¶Õ
domain frequency domain
â´àÁ¹àÇÅÒ
time domain
µÑǤٳÅÒ¡ÃÒ¹¨ì µÑÇªÕ ºÍ¡
domain
Lagrange multiplier
indicator
µÑÇ´Óà¹Ô¹¡Òä͹âÇÅ٪ѹ µÑÇ´Óà¹Ô¹¡ÒäèÒ¤Ò´ËÁÒ µÑÇ´Óà¹Ô¹¡ÒÃ˹èǧàÇÅÒ µÑǹÓ˹éÒ µÒÃÒ§¤é¹ËÒ
convolution operator expectation operator delay operator
predecessor look up table
183
·ÒÃìà¡çµ
target
·ÒÃìà¡çµ·Õ àËÁÒÐ·Õ ÊØ´
optimal target
·ÒÃìà¡çµáºº GPR
generalized partial response (GPR) target
·ÒÃìà¡çµáºº PR
partial response (PR) target
·ÓãËéà» ¹ºÃ÷Ѵ°Ò¹
normalize
à·¤¹Ô¤¡ÒûÃÐÁÒ³¤èÒ㹪èǧ à·ÃÐ亵ì á·ç»
interpolation technique
TB (terabyte) tap
ä·ÁÁÔ §¿ §¡ìªÑ¹
timing function
ä·ÁÁÔ §¿ §¡ìªÑ¹áºº¹ÍÃìÁÍÅäÅ«ì
normalized timing function
timing recovery
ä·ÁÁÔ §ÃԤѿàÇÍÃÕ
ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Õ ãªé¡Ñ¹·Ñ Çä» ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ¹ÔùÑÂ
conventional timing recovery
deductive timing recovery
ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ»ÃÐÁÒ³¤èÒ㹪èǧ ä·ÁÁÔ §ÃԤѿàÇÍÃÕẺÍØ»¹Ñ ä·ÁÁÔ §ÅÙ» ºÔµ
interpolated timing recovery
inductive timing recovery
timing loop
bit
ºÔµ¢èÒÇÊÒÃ
message bit
ºÔµà«ÅÅì (¤Òº¢Í§àÇÅÒã¹Ë¹Ö §ºÔµ) ºÔµà»ÅÕ Â¹Ê¶Ò¹Ð ºÔµÊèǹà¡Ô¹ ẹ´ìÇÔ´·ì
bit cell
transition bit
redundant bit bandwidth
ẹ´ìÇÔ´·ìà¡Ô¹à» ¹ÈÙ¹Âì ẹ´ìÇÔ´·ì¢Í§ÅÙ»
zero excess bandwidth
loop bandwidth
Ẻ¢éÍÁÙÅ
data pattern
Ẻ¨ÓÅͧ
model
Ẻ¨ÓÅͧ¡Ò÷ӧҹ¢Í§Ç§¨Ãà¿ÊÅçÍ¡ÅٻẺàªÔ§àÊé¹ áºº¨ÓÅͧªèͧÊÑÒ³·Õ äÁèµèÍà¹× ͧ·Ò§àÇÅÒẺÊÁÁÙŠẺ¨ÓÅͧªèͧÊÑÒ³àÊÁ×͹¨ÃÔ§ Ẻ¨ÓÅͧªèͧÊÑÒ³ÍØ´Á¤µÔ
亹ÒÃÕ (°Ò¹Êͧ)
ideal channel model
binary
»ÃÐÊÔ·¸Ô¼Å¢Í§ÃËÑÊ
equivalent discrete time channel model
realistic channel model
byte
亵ì (1 亵ì = 8 ºÔµ)
linearized phase locked loop (PLL) model
code e ciency
»ÃѺ¤èÒ (ãËéà» ¹» ¨¨ØºÑ¹)
update
»ÃÔÁҳ˹èǧàÇÅÒã¹ÅÙ»
loop delay
184
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
¼ÅµÍºÊ¹Í§
response
¼ÅµÍºÊ¹Í§¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð ¼ÅµÍºÊ¹Í§¢Í§Ãкº
system response
¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ ¼ÅµÍºÊ¹Í§ä´ºÔµ
transition response
frequency response dibit response
¼ÅµÍºÊ¹Í§·ÒÃìà¡çµ
target response
¼ÅµÍºÊ¹Í§ºÒ§Êèǹ
PR (partial response) PRML (partial response maximum likelihood)
¼ÅµÍºÊ¹Í§ºÒ§Êèǹ¤ÇèÐà» ¹ÁÒ¡ÊØ´ ¼ÅµÍºÊ¹Í§ºÒ§ÊèǹẺ·Ñ Çä» ¼ÅµÍºÊ¹Í§ÍÔÁ¾ÑÅÊì
impulse response FIR ( nite impulse response)
¼ÅµÍºÊ¹Í§ÍÔÁ¾ÑÅÊì¨Ó¡Ñ´ ¼ÅµÍºÊ¹Í§ÍÔÁ¾ÑÅÊìäÁè¨Ó¡Ñ´ á¼è¹«Õ´Õ
IIR (in nite impulse response)
CD (compact disc)
á¼è¹´ÕÇÕ´Õ
DVD (digital versatile disc)
á¼è¹ºÑ¹·Ö¡áÁèàËÅç¡ á¼¹ÀÒ¾
magnetic oppy disk
diagram
á¼¹ÀÒ¾à·ÃÅÅÔÊ ¾Åѧ§Ò¹
GPR (generalized partial response)
trellis diagram
energy
¾Åѧ§Ò¹Ë¹Ö §Ë¹èÇÂ
unit energy
¾ËعÒÁ (â¾ÅÔâ¹àÁÕÂÅ)
polynomial
à¾Íà«ÍÃìäÇàÇÍÃìä·ÁÁÔ §ÃԤѿàÇÍÃÕ
per survivor timing recovery
à¾Íà«ÍÃìäÇàÇÍÃìä·ÁÁÔ §ÃԤѿàÇÍÃÕẺ·Ó§Ò¹« Ó â¾Å ¿ÅÑ¡«ì ¿ §¡ìªÑ¹
per survivor iterative timing recovery
pole ux function
¿ §¡ìªÑ¹¢Ñ ¹ºÑ¹ä´
step function
¿ §¡ìªÑ¹¤ÇÒÁ˹Òá¹è¹¤ÇÒÁ¹èÒ¨Ðà» ¹
probability density function
¿ §¡ìªÑ¹¤ÇÒÁ˹Òá¹è¹¤ÇÒÁ¹èÒ¨Ðà» ¹áººà¡ÒÊìà«Õ¹ ¿ §¡ìªÑ¹â¤Ã๤à¡ÍÃìà´ÅµÒ ¿ §¡ìªÑ¹ä´àäà´ÅµÒ
Kronecker delta function
Dirac delta function
¿ §¡ìªÑ¹¶èÒÂâ͹
transfer function
¿ §¡ìªÑ¹ÍÔÁ¾ÑÅÊì
impulse function
ä¿¿ Ò¡ÃÐáʵç ÀÒÇÐ
d.c. (direct current)
mode
ÀÒÇСÒÃä´éÁÒ
Gaussian probability density function
acquisition mode
185
ÀÒÇСÒÃä´éÁÒẺÊÁºÙóì ÀÒÇСÒõԴµÒÁ
perfect acquisition
tracking mode
ÀÒÇСÒý ¡ÍºÃÁ
training mode
àÁµÃÔ¡ (µÑÇÇÑ´)
metric
àÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð àÁ·ÃÔ¡«ìÊËÊÑÁ¾Ñ¹¸ì¢éÒÁ
cross correlation matrix
àÁ·ÃÔ¡«ìÍѵÊËÊÑÁ¾Ñ¹¸ì
auto correlation matrix identity matrix
àÁ·ÃÔ¡«ìàÍ¡Åѡɳì äÁèà¢éҨѧËÇÐ
state transition matrix
asynchronous
äÁèÁÕÊËÊÑÁ¾Ñ¹¸ì¡Ñ¹
uncorrelated ECC (error correction code)
ÃËÑÊá¡é䢢éͼԴ¾ÅÒ´ ÃËÑʤ͹âÇÅ٪ѹ ÃËÑÊÁÍ´ÙàŪѹ
convolutional code modulation code
ÃдѺ¢Õ´àÃÔ Áà»ÅÕ Â¹
threshold level
Ãкº¡ÒèѴà¡çº¢éÍÁÙÅ
data storage system
Ãкº¡Òúѹ·Ö¡áÁèàËÅç¡ Ãкº·Õ ¶Ù¡à¢éÒÃËÑÊ
magnetic recording system
coded system
Ãкº·Õ ÁÕᶺ¤ÇÒÁ¶Õ ¨Ó¡Ñ´ Ãкº·Õ äÁèä´é¶Ù¡à¢éÒÃËÑÊ ÃкºÊ× ÍÊÒôԨԷÑÅ ÃÐÂÐ
band limited system uncoded system
digital communication system
stage
ÃÐÂзҧ¡ÓÅѧÊͧà©ÅÕ Â
MSD (mean squared distance)
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å
e ective distance
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧÊͧ ÃÐÂзҧÂؤÅÔ´
Euclidean distance
ÃÐÂзҧÂؤÅÔ´¡ÓÅѧÊͧ ÃÒ¡¡ÓÅѧÊͧà©ÅÕ Â ÃÙ»¤Å× ¹
squared e ective distance
squared Euclidean distance
RMS (root mean square)
waveform
àèÔÊàµÍÃìẺàÅ× Í¹
shift register
ÅÍ¡ÒÃÔ·ÖÁ¸ÃÃÁªÒµÔ
natural logarithm
Åͧ¼Ô´Åͧ¶Ù¡ ÅӴѺ
trial and error
sequence
ÅӴѺ¢éͼԴ¾ÅÒ´
error sequence
ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ
input error sequence
ÅӴѺ¢éͼԴ¾ÅÒ´ÍÔ¹¾Øµ·Õ ¶Ù¡µéͧ
valid input error sequence
186
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ÅӴѺ¢éÍÁÙÅ
data sequence
ÅӴѺ¢éÍÁÙÅÂèÍÂàÅ¢¤Õ
odd subsequence
ÅӴѺ¢éÍÁÙÅÂèÍÂàÅ¢¤Ùè
even subsequence
ǧ¨Ã VCO ǧ¨Ã¡Ãͧ
VCO (voltage controlled oscillator) lter interpolation lter
ǧ¨Ã¡Ãͧ¡ÒûÃÐÁÒ³¤èÒ㹪èǧ ǧ¨Ã¡Ãͧ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒ ǧ¨Ã¡Ãͧ·Ó¹ÒÂ
prediction error lter
predictor lter
ǧ¨Ã¡Ãͧ㹡Ò÷ÓãËéÊÑҳú¡Ç¹à» ¹ÊÕ¢ÒÇ Ç§¨Ã¡ÃͧẺàªÔ§àÊé¹
linear lter LPF (low pass lter)
ǧ¨Ã¡Ãͧ¼èÒ¹µ Ó
ǧ¨Ã¡Ãͧ¼èÒ¹µ ÓÍØ´Á¤µÔ
ideal low pass lter
loop lter
ǧ¨Ã¡ÃͧÅÙ» ǧ¨Ãà¢éÒÃËÑÊ
encoder
ǧ¨Ãà¢éÒÃËÑÊá¡é䢢éͼԴ¾ÅÒ´ ǧ¨Ãà¢éÒÃËÑÊÁÍ´ÙàŪѹ ǧ¨ÃªÑ¡µÑÇÍÂèÒ§ ǧ¨ÃµÃǨËÒ
noise whitening lter
error correction code (ECC) encoder
modulation encoder
sampler detector
ǧ¨ÃµÃǨËÒ¢éͼԴ¾ÅÒ´·Ò§àÇÅÒ
TED (timing error detector)
ǧ¨ÃµÃǨËÒ¢Õ´àÃÔ Áà»ÅÕ Â¹·Õ äÁèÁÕ˹èǤÇÒÁ¨Ó
memoryless threshold detector
ǧ¨ÃµÃǨËÒ¢Õ´àÃÔ Áà»ÅÕ Â¹·Õ äÁèÁÕ˹èǤÇÒÁ¨ÓẺËÅÒÂÃдѺ
multi level memoryless threshold de
tector ǧ¨ÃµÃǨËÒ¢Õ´àÃÔ Áà»ÅÕ Â¹áººËÅÒÂÃдѺ ǧ¨ÃµÃǨËҨشÊÙ§ÊØ´ ǧ¨ÃµÃǨËÒÅӴѺ
peak detector sequence detector
ǧ¨ÃµÃǨËÒÅӴѺ·Õ ¤ÇèÐà» ¹ÁÒ¡ÊØ´ ǧ¨ÃµÃǨËÒÅӴѺ·Õ àËÁÒÐ·Õ ÊØ´ ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ
MLSD (maximum likelihood sequence detector)
optimal sequence detector
Viterbi detector
ǧ¨ÃµÃǨËÒÊÑÅѡɳì ǧ¨ÃµÃǨËÒ·Õ àËÁÒÐ·Õ ÊØ´ ǧ¨Ã¶Í´ÃËÑÊ
multi level threshold detector
symbol detector opimal detector
decoder
ǧ¨Ã¶Í´ÃËÑÊá¡é䢢éͼԴ¾ÅÒ´ ǧ¨Ã¶Í´ÃËÑÊÁÍ´ÙàŪѹ
error correction code (ECC) decoder
modulation decoder
ǧ¨Ã¡Ãͧ·Ó¹ÒÂË¹Ö §¢Ñ ¹áººàªÔ§àÊé¹
linear one step predictor
ǧ¨Ãà»ÅÕ Â¹ÊÑÒ³á͹ÐÅçÍ¡à» ¹ÊÑÒ³´Ô¨Ô·ÑÅ
ADC (analog to digital converter)
187
ǧ¨Ãà¿ÊÅçÍ¡ÅÙ»
PLL (phase locked loop)
ǧ¨ÃÀÒ¤ÃѺ
receiver
ǧ¨ÃÀÒ¤Êè§
transmitter
ǧ¨ÃÁÍ´ÙàÅàµÍÃì
modulator
ǧ¨ÃËÒ͹ؾѹ¸ì
di erentiator
àÇ¡àµÍÃìÅѡɳÐ੾ÒÐ
eigenvector
àÇ¡àµÍÃìÅѡɳÐ੾ÒÐẺ¹ÍÃìÁÍÅäÅ«ì stationary
Ê൪ѹà¹ÃÕ Ê¶Ò¹Ð
normalized eigenvector
state
ʶҹеèÍä»
next state start state
ʶҹÐàÃÔ Áµé¹
ʶһ µÂ¡ÃÃÁªèͧÊÑÒ³ÍèÒ¹ Ê໡µÃÑÁ¤èÒÈÙ¹Âì
read channel architecture
spectral null
ÊÀÒ¾¤ÇÒÁà» ¹áÁèàËÅç¡ (¡Ò÷ÓãËéà» ¹áÁèàËÅç¡) ÊÀÒ¾¤ÇÒÁà» ¹áÁèàËÅ硢ͧÊ× ÍºÑ¹·Ö¡ ÊÀҾźÅéÒ§áÁèàËÅç¡ ÊÀÒ¾ãËé«ÖÁ¼èÒ¹ä´é
medium magnetization
coercivity permeability
ÊÁ¡ÒùÍÃìÁÍÅ
normal equation
ÊÁ¡ÒûÃѺ¤èÒ
update equation
ÊÅѺà»ÅÕ Â¹ (·ÃÒ¹Êâ¾Ê) Êèǹ»ÃСͺ
transpose
component
ÊËÊÑÁ¾Ñ¹¸ì
correlation
ÊËÊÑÁ¾Ñ¹¸ì¢éÒÁ ÊÑÒ³
magnetization
cross correlation
signal
ÊÑҳ亹ÒÃÕ (ÊÑÒ³ÊͧÃдѺ) ÊÑÒ³¾ÑÅÊìä´ºÔµ
dibit pulse
ÊÑÒ³¾ÑÅÊì乤ÇÔµÊìÍØ´Á¤µÔ ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹Ð
ideal Nyquist pulse transition pulse
ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹ÐàÍ¡à·È ÊÑҳú¡Ç¹
binary signal
isolated transition pulse
noise
ÊÑҳú¡Ç¹¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð ÊÑҳú¡Ç¹à¡ÒÊìÊÕ¢ÒÇ
white Gaussian noise
ÊÑҳú¡Ç¹à¡ÒÊìÊÕ¢ÒÇẺºÇ¡ ÊÑҳú¡Ç¹¤ÇÒÁÃé͹
transition noise
AWGN (additive white Gaussian noise)
thermal noise
ÊÑҳú¡Ç¹¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡
media jitter noise
188
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ÊÑҳú¡Ç¹·Õ ¢Ö ¹ÍÂÙè¡Ñºáºº¢éÍÁÙÅ
pattern dependent noise
ÊÑҳú¡Ç¹·Õ ÍÂÙè¹Í¡á¶º¤ÇÒÁ¶Õ ÊÑҳú¡Ç¹áººÊÕ
out of band noise
colored noise
ÊÑҳú¡Ç¹ÊÕ¢ÒÇ
white noise
ÊÑҳú¡Ç¹Ê× ÍºÑ¹·Ö¡
media noise
ÊÑÒ³á͹ÐÅçÍ¡·Ò§ä¿¿ Ò·Õ ä´é¨Ò¡ËÑÇÍèÒ¹ (ÊÑÒ³ read back) coe cient
ÊÑÁ»ÃÐÊÔ·¸Ô
Ê× ÍºÑ¹·Ö¡ (¨Ò¹áÁèàËÅç¡) àʶÕÂÃÀÒ¾
media (or medium)
stable
àÊé¹â¤é§ÃÙ»µÑÇàÍÊ
S curve
àÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè àÊé¹ÊÒ¢Ò
survivor path
branch
˹èǧàÇÅÒ
delay
˹èǤÇÒÁ¨Ó¢Í§ªèͧÊÑÒ³ ˹èǤÇÒÁ¨ÓàÊé¹·Ò§
path memory
ËÅÑ¡¡ÒÃàªÔ§µÑ §©Ò¡
orthogonality principle
write head
ËÑÇà¢Õ¹ ËÑÇÍèÒ¹
channel memory
read head
à˵ءÒóì
event
à˵ءÒóì¢éͼԴ¾ÅÒ´
error event
à˵ءÒóì¢éͼԴ¾ÅÒ´·Õ â´´à´è¹ áËÅ觵鹷ҧ
dominant error event
source
áËÅ觻ÅÒ·ҧ
destination
ÍͿ૵·Ò§¤ÇÒÁ¶Õ
frequency o set
ÍͿ૵·Ò§à¿Ê
phase o set
ÍͿ૵·Ò§à¿Ê¢Í§¡ÒêѡµÑÇÍÂèÒ§ ÍͿ૵·Ò§àÇÅÒ ÍѵÃÒ¡ÒâÂÒÂ
timing o set gain
ÍѵÃÒ¡ÒêѡµÑÇÍÂèÒ§
sampling rate
ÍѵÃÒ¡ÒêѡµÑÇÍÂèҧẺà¡Ô¹¨ÃÔ§ ÍѵÃÒ¡ÒÃÅÙèà¢éÒ
oversampling rate
convergence rate
ÍѵÃÒ¢éͼԴ¾ÅÒ´ºÔµ
BER (bit error rate)
ÍѵÃÒ¢èÒÇÊÒÃàªÔ§àÊ鹡ӡѺ ÍѵÃÒ¤ÇÒÁ˹Òá¹è¹ ÍѵÃÒºÔµ
sampling phase o set
bit rate
asymptotic information rate
density ratio
read back signal
189
ÍѵÃÒÃËÑÊ
code rate
ÍѵÃÒÊ觢éÍÁÙÅ
data rate
ÍѵÃÒÊèǹ¤èÒ¡ÓÅѧà©ÅÕ Â¢Í§ÊÑÒ³·Õ µéͧ¡ÒõèͤèÒ¡ÓÅѧà©ÅÕ Â¢Í§ÊÑҳú¡Ç¹ ratio) ÍѵÊËÊÑÁ¾Ñ¹¸ì
auto correlation
ÍÑÅ¡ÍÃÔ·ÖÁ (¢Ñ ¹µÍ¹ÇÔ¸Õ) ÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ
algorithm
Viterbi algorithm inductive
ÍÔ¹´Ñ¡·Õ¿ (»ÃÐàÀ·¢Í§ËÑÇÍèÒ¹) ÍÔ¹¾Øµ (ÃѺà¢éÒ, ¹Óà¢éÒ) ÍÔ ÁµÑÇ
input
saturated
ÍÕ¤ÇÍäÅ૪ѹẺÊÁºÙóì ÍÕ¤ÇÍäÅà«ÍÃì
perfect equalization
equalizer
ÍÕ¤ÇÍäÅà«ÍÃìẺ PR
partial response (PR) equalizer
ÍÕ¤ÇÍäÅà«ÍÃìẺ¼ÅµÍºÊ¹Í§àµçÁ àÍÒµì¾Øµ (Êè§ÍÍ¡, ¹ÓÍÍ¡) àÍÒµì¾Øµ¢Í§ªèͧÊÑÒ³ á͹ÐÅçÍ¡
àÎÔõ«ì
output channel output
analog
ÎÒÃì´´ÔÊ¡ìä´Ã¿ì
hard disk drive
Hz (hertz)
full response equalizer
SNR (signal to noise
190
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
ºÃóҹءÃÁ
[1] S. X. Wang and A. M. Taratorin, Magnetic information storage technology. San Diego: Academic Press, 1999.
[2] B. Sklar, Digital communications:
fundamentals and applications:
Prentice Hall,
2nd edition, 2001.
[3] A. M. Taratorin, Magnetic recording systems and measurements:
Guzik Technical
Enterprises, 2004.
[4] J. W. M. Bergmans, Digital baseband transmission and recording. Boston/London/ Dordrecht: Kluwer Academic Publishers, 1996.
[5] T. Suzuki, Perpendicular magnetic recording: Its basics and potential for the future, IEEE Trans. on Magnetics, vol. MAG 20, no. 5, pp. 675 680, September 1984.
[6] J. Moon, The role of signal processing in data storage, IEEE Signal Processing Magazine, pp. 54 72, July 1998.
[7] S. B. Wicker, Error control systems for digital communication and storage. New Jersey: Printice Hall International, 1995.
[8] B. Vasic and E. M. Kurtas, Coding and signal processing for magnetic recording systems. New York: CRC Press, 2005. 191
[9] K. A. S Immink, Runlength limited sequences, in Proc. of the IEEE, vol. 78, no. 11, pp. 1745 1759, November 1990.
[10] » ÂÐ â¤ÇÔ¹·ì·ÇÕÇѲ¹ì, ¡ÒûÃÐÁÇżÅÊÑÒ³ÊÓËÃѺ¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅ àÅèÁ 1: ¾× ¹°Ò¹ ªèͧÊÑÒ³ÍèÒ¹ à¢Õ¹,
ÈÙ¹Âìà·¤â¹âÅÂÕÍÔàÅç¡·Ã͹ԡÊìààÅФÍÁ¾ÔÇàµÍÃìààË觪ҵÔ
(๤෤),
2550.
[11] T. A. Roscamp, E. D. Boerner, and G. J. Parker, Three dimensional modeling of perpendicular recording with soft underlayer, J. of Applied Physics, vol. 91, no. 10, May 2002.
[12] A. V. Oppenheim, A. S. Willsky, and S. H. nawab, Signals and systems. New Jersey: Prentice Hall, 2nd edition, 1997.
[13] Available online at http://www.mathworks.com
[14] » ÂÐ â¤ÇÔ¹·ì·ÇÕÇѲ¹ì, ¤ÙèÁ×Íâ»Ãá¡ÃÁÀÒÉÒ SCILAB ÊÓËÃѺ¼ÙéàÃÔ Áµé¹ (¾ÔÁ¾ì¤ÃÑ §·Õ 2), ÈÙ¹Âì¼ÅÔµ µÓÃÒàÃÕ¹, ʶҺѹ෤â¹âÅÂÕ¾ÃШÍÁà¡ÅéÒ¾Ãй¤Ãà˹×Í, 2549.
[15] G. D. Forney, Maximum likelihood sequence estimation of digital sequences in the presence of intersymbol interference, IEEE Trans. on Information Theory, vol. IT 18, no. 3, pp. 363 378, May 1972.
[16] J. R. Barry, E. A. Lee, and D. G. Messerschmitt, Digital communication. Boston: Kluwer Academic Publishers, 3nd edition, 2003.
[17] H. K. Thapar and A. M. Patel, A class of partial response systems for increasing storage density in magnetic recording, IEEE Trans. on Magnetics, vol. 23, no. 5, pp. 3666 3668, September 1987.
[18] P. Kovintavewat, I. Ozgunes, E. Kurtas, J. R. Barry, and S. W. McLaughlin, General ized partial response targets for perpendicular recording with jitter noise, IEEE Trans. on Magnetics, vol. 38, no. 5, pp. 2340 2342, September 2002.
[19] J. Moon and W. Zeng, Equalization for maximum likelihood detector, IEEE Trans. on Magnetics, vol. 31, no. 2, pp. 1083 1088, March 1995.
[20] S. Raghaven and H. K. Thapar, Feed forward timing recovery for digital magnetic recording, in Proc. of ICC 91, vol. 2, pp. 794 798, June 1991.
[21] A. N. D Andrea, U. Mengali, and G. M. Vitetta, Approximate ML decoding of coded PSK with no explicit carrier phase reference, IEEE Trans. on Communication, vol. 42, no. 2/3/4, pp. 1033 1039, Feb/Mar/Apr 1994.
[22] U. Mengali and A. N. D Andrea, Synchronization techniques for digital receivers, New York: Plenum Press, 1997.
[23] J. Moon and L. R. Carley, Performance comparison of detection methods in magnetic recording, IEEE Trans. on Magnetics, vol. 26, pp. 3155 3172, November 1990. [24] K. H. Mueller and M. Müller, Timing recovery in digital synchronous data receivers, IEEE Trans. on Communication, vol. COM 24, no. 5, pp. 516 531, May 1976.
[25] M. H. Hayes, Statistical digital signal processing and modeling. John Wiley & Sons Inc., New York, 1996.
[26] A. Leon Garcia, Probability and random processes for electrical engineering. New York: Addison Wesley Inc., 2nd edition, 1994.
[27] R. D. Cideciyan, F. Dolivo, R. Hermann, W. Hirt, and W. Schott, A PRML system for digital magnetic recording, IEEE J. Selected Areas Commun., vol. 10, no. 1, pp. 38 56, January 1992.
[28] H. Sha ee, Timing recovery for sampling detectors in digital magnetic recording, in Proc. of ICC 96, vol. 1, pp. 577 581, January 1996.
[29] A. R. Nayak, Iterative timing recovery for magnetic recording channels with low signal to noise ratio, Ph.D. dissertation, Georgia Institute of Technology, Georgia, June 2004.
[30] P. Kovintavewat, Timing recovery based on per survivor processing, Ph.D. disserta tion, Georgia Institute of Technology, Georgia, October 2004.
[31] J. R. Barry, A. Kavc̆ić, S. W. McLaughlin, and A. R. Nayak, Iterative timing recovery, IEEE Signal Processing Magazine, vol. 21, pp. 89 102, January 2004.
[32] X. Jin and A. Kavc̆ić, Cycle slip detection using soft output information, in Proc. of ICC 01, vol. 9, pp. 2706 2710, June 2001.
[33] A. R. Nayak, J. R. Barry, and S. W. McLaughlin, Joint timing recovery and turbo equalization for coded partial response channels, IEEE Trans. on Magnetics, vol. 38, no. 5, pp. 2295 2297, September 2003.
[34] F. M. Gardner, Interpolation in digital modems Part I: Fundamentals, IEEE Trans. on Communication, vol. 41, no. 3, pp. 501 507, March 1993.
[35] L. Erup, F. M. Gardner, and R. A. Harris, Interpolation in digital modems Part II: Implementation and performance, IEEE Trans. on Communication, vol. 41, no. 6, pp. 998 1008, June 1993.
[36] Z. Wu, J. M. Cio , and K. D. Fisher A MMSE interpolated timing recovery scheme for the magnetic recording channel, in Proc. of ICC 97, vol. 3, pp. 1625 1629, 1997.
[37] M. Spurbeck and R. T. Behrens, Interpolated timing recovery for hard disk drive read channels, in Proc. of ICC 97, vol. 3, pp. 1618 1624, 1997.
[38] P. Kovintavewat, J. R. Barry, F. M. Erden, and E. M. Kurtas, Per survivor timing recovery for uncoded partial response channels, in Proc. of ICC 04, Paris, France, vol. 5, pp. 2715 2719, June 20 24, 2004.
[39] P. Kovintavewat, J. R. Barry, F. M. Erden, and E. M. Kurtas, Per survivor iterative timing recovery for coded partial response channels, in Proc. of Globecom 04, Texas, USA, vol. 4, pp. 2604 2608, Nov 29 Dec 3, 2004.
[40] P. Kovintavewat, J. R. Barry, M. F. Erden, and E. M. Kurtas, Robustness of Per Survivor Iterative Timing Recovery in Perpendicular Recording Channels, IEEE In ternational Conference on Magnetics (INTERMAG 2005), Nagoya, Japan, pp. 1613 1614, April 4 8, 2005.
[41] D. G. Messerschmitt, Design of nite impulse response for the Viterbi algorithm and decision feedback equalizer, in Proc. of ICC 74, pp. 37D 1 5, June 1974.
[42] J. Fitzpatrick, J. K. Wolf, and L. Barbosa, New equalizer targets for sampled magnetic recording system, in Proc. of the 25th Asilomar Conference on Signals Systems and Computers, pp. 30 34, November 1991.
[43] T. Oenning and J. Moon, Partial response maximum likelihood detection for perpen dicular recording, IEEE International Conference on Magnetics (INTERMAG 2000), p. HT 08, 2000.
[44] I. Lee, C. Modlin, and J. M. Cio , Equalized maximum likelihood receiver in a magnetic recording channel, in Proc. of ICC 93, pp. 1970 1973, November 1993.
[45] C. T. Beare, The choice of the desired impulse response in combined linear Viterbi algorithm equalizers, IEEE Trans. on Communication, vol. COM 26, no. 8, pp. 1301 1307, August 1978.
[46] I. Lee and J. M. Cio , Equalized maximum likelihood receiver with a unit energy constraint, IEEE Trans. on Magnetics, vol. 33, no. 1, pp. 855 862, January 1997.
[47] A. Ghrayeb and W. E. Ryan, Precoder design for concatenating convolutional codes with generalized partial response channels, in Proc. of Globecom 00, San Francisco, CA, 2000, pp. 1859 1864.
[48] L. C. Barbosa, Maximum likelihood sequence estimators: a geometric view, IEEE Trans. on Information Theory, vol. 35, no. 2, pp. 419 427, March 1989.
[49] J. Caroselli and J.
K. Wolf, Error event characterization of partial response systems
in magnetic recording systems with medium noise, in Proc. of Globecom 98, vol. 5, pp. 2724 2728.
[50] B. E. Moision, P. H. Siegel, and E. Soljanin, Distance enhancing for digital recording, IEEE Trans. on Magnetics, vol. 34, no. 1, pp. 68 74, January 1998.
[51] P. R. Chevillat, E. Eleftheriou, and D. Maiwald, Noise predictive partial response equalizers and applications, in Proc. of ICC 92, vol. 2, pp. 942 947, June 1992.
[52] E. Eleftheriou and W. Hirt, Noise predictive maximum likelihood (NPML) detection for the magnetic recording channel, in Proc. of ICC 96, vol. 1, pp. 556 560, June 1996.
[53] S. Altekar, Detection and coding techniques for magnetic recording channels, Ph.D. dissertation, Univ. Calif. San Diego, June 1997.
[54] J. Moon and J. Park, Â&#x201C;Pattern dependent noise prediction in signal dependent noise,Â&#x201D; IEEE J. Selected Areas Comm., vol. 19, no. 4, pp. 730 Â&#x2013; 743, June 2001.
[55] J. Caroselli, S. A. Altekar, P. McEwen, and J. K. Wolf, Â&#x201C;Improved detection for magnetic recording systems with media noise,Â&#x201D; IEEE Trans. on Magnetics, vol. 33, no. 5, pp. 2779 Â&#x2013; 2781, 1997.
[56] X. Yang and E. Kurtas, Â&#x201C;Noise predictive equalization and detection,Â&#x201D; Internal report, Channels Department, Seagate Technology, Pittsburgh, PA, May 2003.
[57] R. Raheli, A. Polydoros, and C. K. Tzou, Â&#x201C;Per survivor processing: a general approach to MLSE in uncertain environments,Â&#x201D; IEEE Trans. on Commun., vol. 43, no. 2, pp. 354 Â&#x2013; 364, Feb./Mar./Apr. 1995.
[58] C. E. Shannon, Â&#x201C;A mathematical theory of communication,Â&#x201D; The Bell system technical journal, vol. 27, pp. 379 Â&#x2013; 423, 623 Â&#x2013; 656, July, October, 1948.
[59] P. H. Siegel and J. K. Wolf, Â&#x201C;Modulation and coding for information storage,Â&#x201D; IEEE Communications Magazine, pp. 68 Â&#x2013; 86, December 1991.
[60] K. A. S. Abdel Gha ar and J. H. Weber, Â&#x201C;Constrained block codes for class IV partial response channels with maximum likelihood sequence estimation,Â&#x201D; IEEE Trans. on Information Theory, vol. 42, no. 5, pp. 1405 Â&#x2013; 1424, September 1996.
´Ãê¹Õ PW50 ,
6
ËÅÑ¡¡ÒÃàªÔ§µÑ §©Ò¡, 142
¡Òúѹ·Ö¡áººá¹ÇµÑ §, 6
ÍÑÅ¡ÍÃÔ·ÖÁ, 140 144
¡Òúѹ·Ö¡áººá¹Ç¹Í¹, 6
à¾Íà«ÍÃìäÇàÇÍÃì (PSP), 144 àÁµÃÔ¡ÊÒ¢Ò, 140, 141
GPRML, 133
PLL, 18, 21, 67 NPML, 115 135, 138
¡ÒÃÍ͡Ẻ¤èÒ¾ÒÃÒÁÔàµÍÃì, 23 30
¡Ò÷ÓãËéÊÑҳú¡Ç¹à» ¹ÊÕ¢ÒÇ, 117
Íѹ´Ñº·Õ Êͧ
¢éͼԴ¾ÅÒ´¡ÓÅѧÊͧà©ÅÕ Â, 119
¡ÒÃÇÔà¤ÃÒÐËìàªÔ§àÊé¹, 28 30
·Õ ¹éÍÂÊØ´, 120
ÊÁ¡ÒûÃѺ¤èÒ, 22
¤ÇÒÁ«Ñº«é͹, 126
Íѹ´Ñº·Õ Ë¹Ö §
·ÒÃìà¡çµ»ÃÐÊÔ·¸Ô¼Å, 122, 123
¡ÒÃÇÔà¤ÃÒÐËìàªÔ§àÊé¹, 23 28
»ÃÐÊÔ·¸ÔÀÒ¾, 129
ÊÁ¡ÒûÃѺ¤èÒ, 22
ǧ¨Ã¡Ãͧ·Ó¹ÒÂ, 117, 118 ËÅÑ¡¡Ò÷ӧҹ, 120 129 ËÅÑ¡¡ÒÃàªÔ§µÑ §©Ò¡, 119
àʶÕÂÃÀÒ¾, 24, 28 preamble, 22, 34 PRML, 44, 65 88, 115, 126, 137
àÁµÃÔ¡ÊÒ¢Ò, 121, 122
ÃËÑÊ
(0, G/I),
166
ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ, 69
PDNP, 137 150 ¡Ò÷ÓãËéÊÑҳú¡Ç¹à» ¹ÊÕ¢ÒÇ, 141
ÍÑÅ¡ÍÃÔ·ÖÁÇÕà·ÍÃìºÔ, 75
¤ÇÒÁ«Ñº«é͹, 146
ÍÕ¤ÇÍäÅà«ÍÃì, 66
»ÃÐÊÔ·¸ÔÀÒ¾, 146
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´, 71
ǧ¨Ã¡Ãͧ·Ó¹ÒÂ, 141
á¼¹ÀÒ¾à·ÃÅÅÔÊ, 72 199
SCILAB, 8, 54
ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹Ð, 6
SNR, 21, 104, 130
à˵ءÒóì¢éͼԴ¾ÅÒ´, 96
Electronics, 55
¡Òúѹ·Ö¡áººäº¹ÒÃÕ, 5
¢Í§à˵ءÒóì¢éͼԴ¾ÅÒ´, 100
¤ÇÒÁäÁèà» ¹àªÔ§àÊé¹, 5
µèͺԵ (Eb /N0 ), 31
¡ÒûÃѺ¤èÒ·Ò§àÇÅÒ, 37
»ÃÐÊÔ·¸Ô¼Å, 47, 101, 103
¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ, 47 54 Ẻ
¡®¢Í§¿ÒÃÒà´Âì, 5
Ẻ¾Åѧ§Ò¹Ë¹Ö §Ë¹èÇÂ, 53
¡Ò÷ÓãËéÊÑҳú¡Ç¹à» ¹ÊÕ¢ÒÇ, 116, 138,
ÍèÒ¹, 5 8 à¢Õ¹, 5 ¡ÃÐáÊä¿¿ Òà¢Õ¹, 4, 5, 152 ¡ÒáŠÓÃËÑʾÑÅÊì (PCM), 2 ¡ÒêѡµÑÇÍÂèÒ§, 11, 18, 21, 36 ¡Òúǡ ¡ÒÃà»ÃÕºà·Õº ¡ÒÃàÅ×Í¡, 122 ¡Òúѹ·Ö¡ÃкºáÁèàËÅç¡, 1 Ẻ¨ÓÅͧªèͧÊÑÒ³, 9 ¡Òúѹ·Ö¡áººá¹ÇµÑ §, 3, 44, 54, 96, 103, 147
ẺâÁ¹Ô¡ (h0
¡Òúѹ·Ö¡áººá¹Ç¹Í¹, 3, 96, 129, 147
49
ẺÊÁºÙóì, 55, 104, 130, 147 ¡ÒÃà´Ô¹áººÊØèÁ, 20 ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð, 4, 11, 33, 76, 139, 152 ·Õ ´Õ·Õ ÊØ´, 77 ¼ÅµÍºÊ¹Í§, 6, 45 ÊÀÒ¾¤ÇÒÁà» ¹áÁèàËÅç¡, 5 àÁ·ÃÔ¡«ì, 158 àÍ¡à·È, 6, 55, 104, 130 ¡ÒÃàÅ× Í¹µÓá˹觢ͧ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹ÐẺÊØèÁ, 55, 139 ¡ÒÃá·Ã¡ÊÍ´ÃÐËÇèÒ§ÊÑÅѡɳì (ISI), 6, 65,
ÊÑÒ³¾ÑÅÊìà»ÅÕ Â¹Ê¶Ò¹Ð, 6 à˵ءÒóì¢éͼԴ¾ÅÒ´, 97
= 1),
¡ÒÃà¢éҨѧËÇÐ, 18, 22, 34, 67
·ÒÃìà¡çµáºº PR, 13, 44 ¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ , 8
52
Ẻ·ÒÃìà¡çµà©¾ÒÐ, 53
¡Ãкǹ¡ÒÃ
141
h1 = 1,
69, 152 ¡ÒÃá»Å§
Z,
24, 28
¡ÒÃá»Å§¿ÙàÃÕÂÃì·Õ µèÍà¹× ͧ·Ò§àÇÅÒ, 8
·ÒÃìà¡çµáºº PR, 12, 44 ¼ÅµÍºÊ¹Í§àªÔ§¤ÇÒÁ¶Õ , 8
¢Õ´¨Ó¡Ñ´«Ùà»ÍÃì¾ÒÃÒáÁ¡à¹µÔ¡, 3
¢éͼԴ¾ÅÒ´
¡Òúѹ·Ö¡áººá¹Ç¹Í¹, 44
¡ÓÅѧÊͧà©ÅÕ Â (MSE), 47, 119 ºÔµà«ÅÅì, 6, 24, 55, 103, 147 ¡ÓÅѧÊͧà©ÅÕ Â·Õ ¹éÍÂÊØ´ (MMSE), 47, 120 ¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ, 47 54 ·Ò§àÇÅÒ, 21, 24, 30, 38 ẺÃÒ¡¡ÓÅѧÊͧà©ÅÕ Â (RMS), 34
¼ÅµÍºÊ¹Í§, 116 ¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð, 6, 45 ¢Ñ ¹ºÑ¹ä´, 25 ¢éͼԴ¾ÅÒ´, 26
¤ÇèÐà» ¹ÁÒ¡ÊØ´ (ML), 44 ¤ÇÒÁ¶Õ Ẻ¹ÍÃìÁÍÅäÅ«ì, 8 ¤ÇÒÁ¶Õ 乤ÇÔµÊì, 44
·ÒÃìà¡çµ, 12, 45, 68, 138 ºÒ§Êèǹ (PR), 12, 43, 68 ÍÔÁ¾ÑÅÊì, 20, 30, 148
¤ÇÒÁ˹Òá¹è¹
ÍÔÁ¾ÑÅÊì¨Ó¡Ñ´ (FIR), 67, 117, 138
¡ÒúÃèØ, 155
ÍÔÁ¾ÑÅÊìäÁè¨Ó¡Ñ´ (IIR), 117
¢Í§¡Òúѹ·Ö¡¢éÍÁÙÅ, 6 ¢Í§¡Òúѹ·Ö¡áºº¹ÍÃìÁÍÅäÅ«ì (ND), 6 Ê໡µÃÑÁ¡ÓÅѧ, 20, 55, 103, 130, 147
ªèͧÊÑÒ³ÍèÒ¹, 4
àªÔ§¤ÇÒÁ¶Õ , 8, 43 àµçÁ, 67 ä´ºÔµ, 8, 45
¿ §¡ìªÑ¹ ¢Ñ ¹ºÑ¹ä´, 23
·ÒÃìà¡çµ, 53
¢éͼԴ¾ÅÒ´, 6
·Õ àËÁÒÐ·Õ ÊØ´, 47 »ÃÐÊÔ·¸Ô¼Å, 122
¤ÇÒÁ˹Òá¹è¹¤ÇÒÁ¹èÒ¨Ðà» ¹áººà¡ÒÊìà«Õ¹, 55, 78, 103, 139, 147
Ẻ GPR, 13, 45 Ẻ
h1 = 1,
52
«Ô§¡ì, 20 ¶èÒÂâ͹, 13, 24, 46, 117, 138
Ẻ·ÒÃìà¡çµà©¾ÒÐ, 53 àÅ¢ªÕ ¡ÓÅѧ, 8 Ẻ¾Åѧ§Ò¹Ë¹Ö §Ë¹èÇÂ, 53 â¤Ã๤à¡ÍÃìà´ÅµÒ, 81 ẺâÁ¹Ô¡, 49, 104, 133, 148 Ẻ PR, 12 ¡Òúѹ·Ö¡áººá¹ÇµÑ §, 44
ÀÒÇÐ, 22 ¡ÒõԴµÒÁ, 22
Ãкº¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅ, 3
¡ÒÃä´éÁÒ, 22 ¡ÒÃä´éÁÒẺÊÁºÙóì, 34
Ẻ¨ÓÅͧ, 4 ÃкºÊ× ÍÊÒôԨԷÑÅ, 17
ÃËÑÊ
(0, G/I),
166 167
ÃٻẺ NRZI, 4, 151
ÃËÑÊ RLL, 3, 91, 151 167 ÅӴѺ¢éͼԴ¾ÅÒ´, 93
ISI, 152 ¡ÒÃÍ͡Ẻ, 162 164 ¤ÇÒÁ˹Òá¹è¹¡ÒúÃèØ, 155 ¤ÇÒÁ¨Ø, 154 155 µÒÃÒ§¤é¹ËÒ, 152, 164
ÍÔ¹¾Øµ·Õ ¶Ù¡µéͧ, 94
ǧ¨Ã VCO, 18 ǧ¨Ã¡Ãͧ ¢éͼԴ¾ÅÒ´¡Ò÷ӹÒÂ, 118
ºÔµÊèǹà¡Ô¹, 153
·Ó¹ÒÂ, 117
¾ÒÃÒÁÔàµÍÃì, 151
·Ó¹ÒÂË¹Ö §¢Ñ ¹áººàªÔ§àÊé¹, 118
ÃËÑÊ
(0, G/I),
166
ÃËÑÊ FM, 164 ÃËÑÊ MFM, 164
¼èÒ¹µ Ó, 4 ¤ÇÒÁ¶Õ µÑ´, 20 ÅÙ», 18
ÃËÑÊ Miller, 164
ǧ¨ÃªÑ¡µÑÇÍÂèÒ§, 4, 18
ÍѵÃÒ¢èÒÇÊÒÃàªÔ§àÊ鹡ӡѺ, 155, 160
ǧ¨ÃµÃǨËÒ, 4
ÍѵÃÒ¤ÇÒÁ˹Òá¹è¹, 155 156
¢Õ´àÃÔ Áà»ÅÕ Â¹
ÍѵÃÒÃËÑÊ, 152
·Õ äÁèÁÕ˹èǤÇÒÁ¨Ó, 31
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´, 156 158
·Õ äÁèÁÕ˹èǤÇÒÁ¨ÓẺËÅÒÂÃдѺ, 34
à§× ͹䢺ѧ¤Ñº
ẺËÅÒÂÃдѺ, 21
¨Ó¹Ç¹ÅӴѺ¢éÍÁÙÅ, 153 154 àÁ·ÃÔ¡«ì¡ÒÃà»ÅÕ Â¹Ê¶Ò¹Ð, 158 159 ä·ÁÁÔ §ÃԤѿàÇÍÐÃÕ, 152
ẺÍè͹, 38 Ẻá¢ç§, 38 ¢éͼԴ¾ÅÒ´·Ò§àÇÅÒ, 18
ÃËÑÊ RS, 3
·Õ àËÁÒÐ·Õ ÊØ´, 69
ÃËÑʤ͹âÇÅ٪ѹ, 69
ÅӴѺ·Õ ¤ÇèÐà» ¹ÁÒ¡ÊØ´ (MLSD), 77
ÃËÑÊÁÍ´ÙàŪѹ, 151
ÅӴѺàËÁÒÐ·Õ ÊØ´, 137
ÇÕà·ÍÃìºÔ, 12, 21, 69, 115 ¤ÇÒÁ«Ñº«é͹, 80
¡Òúѹ·Ö¡áººá¹Ç¹Í¹, 6 àÍ¡à·È, 55, 104, 130
¤ÇÒÁÅÖ¡¡ÒöʹÃËÑÊ, 80
ä´ºÔµ, 8
¤ÇÒÁ¹èÒ¨Ðà» ¹¢Í§¢éͼԴ¾ÅÒ´ÅӴѺ, 88
乤ÇÔµÊì, 20
»ÃÐÊÔ·¸ÔÀÒ¾, 87
乤ÇÔµÊìÍØ´Á¤µÔ, 12
ǧ¨ÃµÃǨËÒÅӴѺ, 69
ÊÑҳú¡Ç¹, 11, 20, 21, 45, 68, 93
ÍÑÅ¡ÍÃÔ·ÖÁ, 69, 75 80
¡ÒâÂÒÂ, 45, 66
à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´, 70 72
¡ÒÃ¢Ö ¹ÍÂÙè¡Ñºáºº¢éÍÁÙÅ, 138 140
á¼¹ÀÒ¾à·ÃÅÅÔÊ, 70, 72 75
¡Ò÷ӹÒÂ, 117, 141
ÊÑÅѡɳì, 21
¨ÔµàµÍÃì¢Í§Ê× ÍºÑ¹·Ö¡, 55, 61, 103, 107, 138, 139
ǧ¨Ã¶Í´ÃËÑÊ ÁÍ´ÙàŪѹ, 4
·Õ ¢Ö ¹ÍÂÙè¡Ñºáºº¢éÍÁÙÅ, 138
á¡é䢢éͼԴ¾ÅÒ´, 4
·Õ ÍÂÙè¹Í¡á¶º¤ÇÒÁ¶Õ , 11, 20
ǧ¨ÃÁÍ´ÙàÅàµÍÃì, 4, 5
ÀÒÂã¹Ç§¨ÃªÑ¡µÑÇÍÂèÒ§, 18
ǧ¨Ãà¢éÒÃËÑÊ
ÊÕ¢ÒÇ, 60, 119
¡è͹, 91 ÁÍ´ÙàŪѹ, 3
à¡ÒÊìÊÕ¢ÒÇ, 69 à¡ÒÊìÊÕ ¢ÒÇẺºÇ¡, 20, 55, 66, 75, 99, 103, 115, 138
á¡é䢢éͼԴ¾ÅÒ´, 3
ẺÊÕ, 93, 99, 116, 138 ʶһ µÂ¡ÃÃÁªèͧÊÑÒ³ÍèÒ¹, 11
Ê໡µÃÑÁ¤èÒÈÙ¹Âì, 8, 44, 68
ÊÀÒ¾¤ÇÒÁà» ¹áÁèàËÅç¡, 1, 139, 152 ÊÀҾźÅéÒ§áÁèàËÅç¡, 2, 5 ÊÀÒ¾ãËé«ÖÁ¼èÒ¹ä´é, 3
ËÅÑ¡¡ÒÃàªÔ§µÑ §©Ò¡, 119, 142
ÍͿ૵
ÊÁ¡ÒùÍÃìÁÍÅ, 119
·Ò§¤ÇÒÁ¶Õ , 22, 35
ÊÑÒ³¾ÑÅÊì
·Ò§à¿Ê, 21, 22, 33
à»ÅÕ Â¹Ê¶Ò¹Ð, 6, 55, 103, 130, 147 ¡Òúѹ·Ö¡áººá¹ÇµÑ §, 6
ÊÁ¡ÒûÃѺ¤èÒ, 24, 28 ·Ò§àÇÅÒ, 20, 33
ÍѵÃÒ¡ÒêѡµÑÇÍÂèÒ§, 37 Ẻà¡Ô¹¨ÃÔ§, 37
¨Ø´àʶÕÂÃÀÒ¾, 32 à˵ءÒóì¢éͼԴ¾ÅÒ´, 91
ÍѵÃÒ¡ÒÃÅÙèà¢éÒ, 22 25, 28, 38
¡ÒÃÇÔà¤ÃÒÐËì, 91 111
ÍѵÃÒ¢éͼԴ¾ÅÒ´ºÔµ (BER), 47
¡ÒÃÇÔà¤ÃÒÐËìÃÐÂзҧ·Õ ¹éÍÂÊØ´, 104 107
ÍѵÃÒÃËÑÊ, 69, 152, 154, 156
¤ÇÒÁÊÑÁ¾Ñ¹¸ì ÃÐËÇèÒ§
ÍѵÊËÊÑÁ¾Ñ¹¸ì, 49, 60, 100, 119, 142 ÍÑÅ¡ÍÃÔ·ÖÁ
SNReff
áÅÐ BER,
107 109 ¤ÇÒÁËÁÒÂ, 93
PDNP, 140 144
¹ÔÂÒÁ, 95
PS PDNP, 144 146
ÃËÑÊ RLL, 91
ÇÕà·ÍÃìºÔ, 69 80, 122
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å, 99 103
Ẻ»ÃѺµÑÇ, 143
ÃÐÂзҧ»ÃÐÊÔ·¸Ô¼Å¡ÓÅѧÊͧ, 101
ÍÕ¤ÇÍäÅà«ÍÃì, 4, 12, 47, 53, 66, 104
·Õ ¹éÍÂÊØ´, 101
ÍѵÃÒ¡ÒÃÅÙèà¢éÒ, 67
ÃÐÂзҧÂؤÅÔ´, 97 99
Ẻ zero forcing, 138
ÃÐÂзҧÂؤÅÔ´¡ÓÅѧÊͧ, 97
Ẻ¼ÅµÍºÊ¹Í§ºÒ§Êèǹ (PR), 43, 68, 115
·Õ ¹éÍÂÊØ´, 99 ǧ¨Ãà¢éÒÃËÑÊ¡è͹, 91
Ẻ¼ÅµÍºÊ¹Í§àµçÁ, 67 ẹ´ìÇÔ´·ì à¤Ã× Í§Ê¶Ò¹Ð¨Ó¡Ñ´, 70, 71 ÃËÑÊ RLL, 156 ǧ¨ÃµÃǨËÒÇÕà·ÍÃìºÔ, 71 à·¤¹Ô¤¡ÒûÃÐÁÒ³¤èÒ㹪èǧ, 37 à¾Íà«ÍÃìäÇàÇÍÃìä·ÁÁÔ §ÃԤѿàÇÍÃÕ, 38 Ẻ·Ó§Ò¹« Ó, 40
¢Í§ÅÙ», 22, 28 à¡Ô¹à» ¹ÈÙ¹Âì, 20 Ẻ¢éÍÁÙÅ, 138 Ẻ¨ÓÅͧ ¡Ò÷ÓãËéÊÑҳú¡Ç¹à» ¹ÊÕ¢ÒÇ, 117 ¡Ò÷ӧҹ¢Í§Ç§¨Ã PLL, 23
àÊé¹·Ò§·Õ ÂѧÁÕªÕÇÔµÍÂÙè, 80, 92
¡ÒÃÍ͡Ẻ·ÒÃìà¡çµ, 47
àÊé¹â¤é§ÃÙ»µÑÇàÍÊ, 21, 30 34
ªèͧÊÑÒ³
¨Ø´ÊÁ´ØÅ, 33
GPR ẺÊÁÁÙÅ, 93
¡Òúѹ·Ö¡ÃкºáÁèàËÅç¡, 9 ·Õ äÁèµèÍà¹× ͧ·Ò§àÇÅÒẺÊÁÁÙÅ, 66 ÍØ´Á¤µÔ, 12 13, 19 àÊÁ×͹¨ÃÔ§, 11 12 Ãкº¡ÒèѴà¡çº¢éÍÁÙÅ´Ô¨Ô·ÑÅ, 3 á¼¹ÀÒ¾à·ÃÅÅÔÊ, 70, 72, 91, 115, 141
ä«à¤ÔÅÊÅÔ», 33 ä·ÁÁÔ §¨ÔµàµÍÃì, 18, 20, 39 ä·ÁÁÔ §¿ §¡ìªÑ¹, 30 Ẻ¹ÍÃìÁÍÅäÅ«ì, 32 ä·ÁÁÔ §ÃԤѿàÇÍÃÕ, 12, 17 41 ¡ÒÃÍ͡Ẻ¤èÒ¾ÒÃÒÁÔàµÍÃì, 23 34 ÀÒÇСÒõԴµÒÁ, 22 ÀÒÇСÒÃä´éÁÒ, 22 Ẻ´Ô¨Ô·ÑÅ, 37 Ẻ·Õ ãªé¡Ñ¹·Ñ Çä», 18 23 »ÃÐÊÔ·¸ÔÀÒ¾, 34 37 Ẻ¹ÔùÑÂ, 18 Ẻ»ÃÐÁÒ³¤èÒ㹪èǧ, 37 ẺÍØ»¹ÑÂ, 18