ENCODED MATTER AADRL 2014 | ROBERT STUART SMITH Andrew Potter Irina Safonova Martina Rosati
ENCODED MATTER The work that follows demonstrates an investigation of CNC milling techniques and their potential ability to receive real time feedback. Through utilization of agent based code, behaviors were modified to achieve varying effects. Working with agent variables such as rotation, alignment, trail attraction, and separation, an understanding of agent interrelations, motions, and navigations was achieved.
ENCODED MATTER
| 1
2 | ENCODED MATTER
TABLE OF CONTENTS 1.0 // MULTIPLE AGENT SIMULATIONS 1.1// FIELD 1.0 1.2// FIELD 2.0 1.3// FIELD 3.0 1.4// FIELD 4.0 1.5// FIELD 5.0 1.6// FIELD 6.0
p. 5 p. 6 p. 10 p. 14 p. 20 p. 26 p. 32
2.0 // SINGLE AGENT SIMULATIONS
p. 37
2.1// SIMULATIONS FOR PATH 2.2// SIMULATIONS FOR TOPOGRAPHY
p. 38 p. 44 p. 48
3.0 // CNC TESTS
p. 50 p. 51 p. 52 p. 53
3.1// TEST 1.0 & TEST 2.0 3.2// TEST 3.0 & TEST 4.0 3.3// TEST 5.0 3.4// TEST 6.0
p. 54
4.0 // REAL TIME MILLING
p. 56 p. 58 p. 60 p. 62 p. 64 p. 66 p. 70 p. 74 p. 78
4.1// FEEDBACK LOOP 4.2// TECHNICAL ISSUES 4.3// TEST 1.0 4.4// TEST 2.0 4.5// TEST 3.0 4.6// TEST 4.0 4.7// TEST 5.0 4.8// TEST 6.0 4.9// TEST 7.0
ENCODED MATTER
| 3
4 | ENCODED MATTER
MULTIPLE AGENT SIMULATIONS Through study of multiple agents the key variables of rotation, range of vision, and attrition to trail. These identified qualities facilitated two modes of investigation: the development of path based simulations and the development of simulations conducive to topography.
PARAMETERS _position _n of Agents _wander _alignment to trail _attraction to trail _alignment _separation _max speed _range of vision _angle of rotation _type transformation
ENCODED MATTER
| 5
6 | ENCODED MATTER
FIELD 1.0 AGENT TRAIL TRANSFORMATION 1 2 3
Type 1 (near trails < 150) Type 2 (150 < near trails < 250) Type 3 (near trails >250)
FIXED RULES Global rules // _position // random Local rules// _attraction to trail // 0.1 _max speed // depending on types _angle of rotation // 10.0 _range of vision // 50.0
VARIABLES _wander _number of agents _alignment to trail
ENCODED MATTER
| 7
1.01 //
1.02 //
Wander // 2.0 Alignment to trail // 1.0 N of Agents // 200
Wander // 1.0 Alignment to trail // 1.0 N of Agents // 200
100
200
300
400
500
8 | ENCODED MATTER
1.03 //
1.04 //
Wander // 2.0 Alignment to trail // 0.5 N of Agents // 100
Wander // 1.0 Alignment to trail // 0.1 N of Agents // 100
100
200
300
400
500
ENCODED MATTER
| 9
10 | ENCODED MATTER
FIELD 2.0 AGENT TRAIL TRANSFORMATION 1 2 3
Type 1 (near trails > 5) && (near trails > 25) Type 2 (near trails > 20) Type 3 (near trails > 75)
FIXED RULES Global rules // _position _number of agents Local rules// _wander // 0.0 _attraction to trail // 1.0 _max speed // depending on types _range of vision // 50.0 _alignment to trail // 0.8
VARIABLES _angle of rotation
ENCODED MATTER |
11
2.01 //
2.02 //
T1 Rotation // -55.0 T2 Rotation // 15.0 T3 Rotation // -15.0
T1 Rotation // -75.0 T2 Rotation // 15.0 T3 Rotation // -10.0
100
200
300
400
500
12 | ENCODED MATTER
2.03 //
2.04 //
T1 Rotation // -75.0 T2 Rotation // 180.0 T3 Rotation // -15.0
T1 Rotation // -75.0 T2 Rotation // 90.0 T3 Rotation // -15.0
100
200
300
400
500
ENCODED MATTER |
13
14 | ENCODED MATTER
FIELD 3.0
FIXED RULES Global rules // _position _number of agents Local rules// _wander // 0.0 _alignment to trail // 0.0 _attraction to trail // 1.0 _separation // 0.0 _max speed // depending on types _range of vision // 80.0 _angle of rotation
VARIABLES 1 2 3 4
Type 1 Type 2 Type 3 Type 4
ENCODED MATTER |
15
3.01 // nearTrail >5// T1 nearTrail >15// T2 nearTrail >20// T2
100
200
300
400
500
16 | ENCODED MATTER
3.02 // nearTrail >25// T4 nearTrail >50// T2
nearTrail >5// T1 nearTrail >15// T2 nearTrail >20// T3
nearTrail >25// T3 nearTrail >50// T2
3.03 // nearTrail >5// T1 nearTrail >15// T2 nearTrail >20// T4
3.04 // nearTrail >25// T3 nearTrail >50// T2
nearTrail >5// T1 nearTrail >15// T2 nearTrail >20// T2
nearTrail >25// T4 nearTrail >50// T1
100
200
300
400
500
ENCODED MATTER |
17
3.01 //
3.02 //
18 | ENCODED MATTER
3.03 //
3.04 //
ENCODED MATTER |
19
20 | ENCODED MATTER
FIELD 4.0 AGENT TRAIL TRANSFORMATION 1 2 3
Type 1 (near trails < 150) Type 2 (150 < near trails < 250) Type 3 (near trails >250)
FIXED RULES Global rules // _position _number of agents Local rules// _wander // 0.0 _attraction to trail // 0.1 _max speed // depending on types _range of vision // 50.0 _angle of rotation // 10.0
VARIABLES _alignment to trail _separation _alignment
ENCODED MATTER |
21
4.01 //
4.02 //
Align to Trail// 0.9 Separation// 1.0 Alignment// 0.0
Align to Trail// 0.5 Separation// 1.5 Alignment// 1.0
40
80
120
160
200
22 | ENCODED MATTER
4.03 //
4.04 //
Align to Trail// 0.8 Separation// 1.0 Alignment// 0.8
Align to Trail// 0.8 Separation// 2.0 Alignment// 0.8
40
80
120
160
200
ENCODED MATTER |
23
4.01 //
4.02 //
24 | ENCODED MATTER
4.03 //
4.04 //
ENCODED MATTER |
25
26 | ENCODED MATTER
FIELD 5.0 AGENT TRAIL TRANSFORMATION 1 2 3
Type 1 (near trails < 150) Type 2 (150 < near trails < 250) Type 3 (near trails >250)
FIXED RULES Global rules // _Position Local rules// _wander // 0.0 _attraction to trail // 0.5 _max speed // depending on types _range of vision // 70.0 _angle of rotation // 10.0
VARIABLES _alignment to trail _n of Agents _range of vision
ENCODED MATTER |
27
5.01 //
5.02 //
Align to trail // 1 N of Agents // 100 Range of Vision // 10
Alignto trail // 1.0 N of Agents // 100 Range of Vision // 20
100
120
140
160
180
28 | ENCODED MATTER
5.03 //
5.04 //
Align to Trail // 0.5 N of Agents // 100 Range of Vision // 50
Align to Trail // 0.5 N of Agents // 200 Range of Vision // 20
100
120
140
160
180
ENCODED MATTER |
29
5.01 //
5.02 //
30 | ENCODED MATTER
5.03 //
5.04 //
ENCODED MATTER |
31
32 | ENCODED MATTER
FIELD 6.0 AGENT TRAIL TRANSFORMATION 1 2 3
Type 1 (near trails < 150) Type 2 (150 < near trails < 250) Type 3 (near trails >250)
FIXED RULES Global rules // _Position Local rules// _wander // 0.0 _attraction to trail // 0.5 _max speed // depending on types _range of vision // 70.0 _angle of rotation // 10.0
VARIABLES _alignment to trail _n of Agents _range of vision
ENCODED MATTER |
33
6.01 //
6.02 //
Range of Vision // 135 T1 Rotation // 10.0 T2 Rotation // -15.0 T3 Rotation // 15.0
Range of Vision // 85
100
120
140
160
180
34 | ENCODED MATTER
T1 Rotation // 10.0 T2 Rotation // -15.0 T3 Rotation // 15.0
6.03 // Range of Vision // 85
6.04 // T1 Rotation // 30.0 T2 Rotation // -15.0 T3 Rotation // -60.0
Range of Vision // 85
T1 Rotation // 30.0 T2 Rotation // -15.0 T3 Rotation // -60.0
100
120
140
160
180
ENCODED MATTER |
35
36 | ENCODED MATTER
SINGLE AGENT SIMULATIONS Simulations for PATH Simulations for TOPOGRAPHY
Through study of multiple agents the key variables of rotation, range of vision, and attraction to trail wer. These identified qualities facilitated two modes of investigation: the development of path based simulations and the development of simulations conducive to topography.
PARAMETERS _position _n of Agents _wander _alignment to trail _attraction to trail _max speed _range of vision _angle of rotation _type transformation
ENCODED MATTER |
37
38 | ENCODED MATTER
SIMULATIONS FOR PATH
PARAMETERS _position _n of Agents _wander _alignment to trail _attraction to trail _max speed _range of vision _angle of rotation _type transformation
ENCODED MATTER |
39
1.01 //
1.02 //
Wander // 0.5 Range of vision // 450
Range of Vision // 100 T1 Rotation // 5.0 T2 Rotation // -15.0 T3 Rotation // 15.0
1000
2000
3000
4000
5000
40 | ENCODED MATTER
1.03 // Range of vision // 50
T1 Rotation // 10.0 T2 Rotation // -15.0 T3 Rotation // -15.0
1000
2000
3000
4000
5000
ENCODED MATTER |
41
1.01 //
1.02 //
Range of Vision // 135 T1 Rotation // 10.0 T2 Rotation // -15.0 T3 Rotation // 15.0
Range of Vision // 85
100
120
140
160
180
42 | ENCODED MATTER
T1 Rotation // 10.0 T2 Rotation // -15.0 T3 Rotation // 15.0
1.03 // Range of Vision // 85
6.04 // T1 Rotation // 30.0 T2 Rotation // -15.0 T3 Rotation // -60.0
Range of Vision // 85
T1 Rotation // 30.0 T2 Rotation // -15.0 T3 Rotation // -60.0
1000
2000
3000
4000
5000
ENCODED MATTER |
43
44 | ENCODED MATTER
SIMULATIONS FOR TOPOGRAPHY
PARAMETERS _position _n of Agents _wander _alignment to trail _attraction to trail _max speed _range of vision _angle of rotation _type transformation
ENCODED MATTER |
45
1.01 //
1.02 //
Range of Vision // 135 T1 Rotation // 10.0 T2 Rotation // -15.0 T3 Rotation // 15.0
Range of Vision // 85
1000
2000
3000
4000
5000
46 | ENCODED MATTER
T1 Rotation // 10.0 T2 Rotation // -15.0 T3 Rotation // 15.0
1.03 // Range of Vision // 85
T1 Rotation // 30.0 T2 Rotation // -15.0 T3 Rotation // -60.0
1000
2000
3000
4000
5000
ENCODED MATTER |
47
48 | ENCODED MATTER
CNC TESTS
METHOD // _ 3D Model _ Greyscale Gradient _ Points Cloud _ RML Real Time
MATERIAL // _blue foam (low density _yellow foam (high density)
MAX DEPTH // BIT // PASSES // MACHINE SETTINGS //
ENCODED MATTER |
49
TEST 1.0 //
TEST 2.0 //
METHOD // 3D Model
METHOD // 3D Model
MATERIAL // Blue Foam
MATERIAL // Blue Foam
MAX DEPTH // 12 mm
MAX DEPTH // 15mm
BIT // a.End Mill 6mm, Ball nose 6 mm b.End Mill 3mm
BIT // End Mill 6mm, VBit 90째
PASSES // a. 2 b. 1
MACHINE SETTINGS // stepover 4mm stepdown 4mm speed 100%
PASSES // 2
MACHINE SETTINGS // a. stepover 3mm stepdown 4mm speed 80%
a.
50 | ENCODED MATTER
b. stepover 1.4mm stepdown 5mm speed 80%
b.
TEST 3.0 //
TEST 4.0 //
METHOD // 3D Model
METHOD // 3D Model
MATERIAL // Blue Foam
MATERIAL // Blue Foam
MAX DEPTH // 15mm
MAX DEPTH // 18mm
BIT // End Mill 6mm | Ball Nose 6 mm
BIT // End Mill 6mm | Ball Nose 6 mm
PASSES // 2
PASSES // 2
MACHINE SETTINGS // Stepover 3mm Stepdown 8mm Speed 20%
Stepover 0.5mm Stepdown 6mm Speed 40%
ENCODED MATTER |
51
TEST 5.0 // METHOD // Greyscale Gradient
MACHINE SETTINGS //
MATERIAL // Yellow Foam
a. Stepdown 2.5 mm Stepover 2 mm Speed 20%
MAX DEPTH // a. 20 mm b. 12 mm c. 15 mm
b. Stepdown 1.5 mm Stepover 2 mm Speed 40%
BIT // a. End Mill 6 mm b. End Mill 6mm | Ball Nose 6 mm c. Ball Nose 6 mm
c. Stepdown 1.5 mm Stepover 1 mm Speed 30%
PASSES // a. 1 ( roughing ) b. 2 ( roughing + finishing ) c. 1 ( finishing
a.
52 | ENCODED MATTER
b.
c.
TEST 6.0 // METHOD // Point Cloud MATERIAL // Yellow Foam
MACHINE SETTINGS // Stepover 3 mm Speed 30%
MAX DEPTH // 15 mm BIT // Ball Nose 6 mm PASSES // 1
ENCODED MATTER |
53
54 | ENCODED MATTER
REAL TIME MILLING
Live feedback testing required the recalibration of previously developed agent behaviors. Through introduction of sensor feedback (a photosensor was chosen for the following investigations ), the influence of agent behavior began either subside or override sensor values. Ultimately, the goal lies somewhere within this fine tuning. Creating pre-established geometries to introduce light and, subsequently, utilizing agent behaviors that achieve a goal based on received values would be considered success.
ENCODED MATTER |
55
FEEDBACK LOOP MACHINE FEEDBACK LOOP // Real Time CNC Feedback loops exist as a continuous flow of responses. When considering the milling process, the introduction of feedback loop attempts to disrupt the typical stream of digital information, whether it be G-Code or RML language, and use actions in progress to shape the final outcome. Through introduction of a photo sensor, the machine is able to respond to varying conditions across the material. This information is sent back to the computer and read again by the CNC, closing the feedback loop.
56 | ENCODED MATTER
FEEDBACK LOOP
PROCESSING Set of local rules in order to obtain varying agent behaviors and produce agent based pattern
OUTPUT Different options for output values
INTENSITY OF LIGHT
LIGHT SOURCE
SENSOR Based on output received from the light source the sensir sends differnet values to the arduino code
ARDUINO TRANSLATE SENSOR VALUES AND SEND TO CNC MACHINE
CNC
SENSOR VALUES EXPRESSED THROUGH DRILL BIT BEHAVIOUR
ENCODED MATTER |
57
TECHNICAL ISSUES ARDUINO AND CNC// METHOD // Mount Arduino and Sensor to CNC PROBLEM // Arduino and sensor cause interference with the CNCâ&#x20AC;&#x2122;s ability to receive computer commands SOLUTION // Laser cut a small attachment that allows the sensor to be mounted exterior to the drillling bit
Arrangment 1
Arduino Board
Arrangment 2
58 | ENCODED MATTER
TECHNICAL ISSUES CNC AND LIGHT SOURCE// METHOD // Create a Light Board PROBLEM // Sensor is not able to interpret light values through full 5.0cm thickness of material SOLUTION // Use larger drilling bits to quickly expose light sources that may be utilized by tools of varying scales
Support for Light Sensor
24mm ball
12mm end
6mm end
6mm ball
3mm end
Light placement on CNC bed
ENCODED MATTER |
59
TEST 1.0 // Original Processing Simulation Torus space becomes a difficulty when attempting to translatet agent simulations into physical manifestations. Bounce was introduced in order to keep the machine within its degined boundaries.
Agent X and Y movements are based on type but influenced by sensor. Rotation and speed change according to light intensity. Z values are determined by sensor readings.
60 | ENCODED MATTER
METHOD // RML real time MATERIAL // Blue Foam MAX DEPTH // 44mm BIT // End Mill 6mm | Ball 24mm PASSES // 2
ENCODED MATTER |
61
TEST 2.0 // CNC AND AGENT// Create an Agent Based Topography RIGHT// Original Processing Simulation Torus space becomes a difficulty when attempting to translatet agent simulations into physical manifestations. Bounce was introduced in order to keep the machine within its degined boundaries.
Agent X and Y movements are based on type but influenced by sensor. Rotation and speed change according to light intensity. Z values are determined by sensor readings.
62 | ENCODED MATTER
METHOD // RML real time MATERIAL // Blue Foam MAX DEPTH // 44mm BIT // End Mill 6mm | Ball 24mm PASSES // 2
ENCODED MATTER |
63
TEST 3.0 // CNC AND AGENT// Create an Agent Based Topography
Original Processing Simulation Agent behaviors conflicted too strongly with the remapped float values. This conflict, in combination with poor sensor readings, made it difficult for any clear geometry to materialize.
X and Y values are based on agent type. Z values are inserted as a float value determined by sensor reading.
64 | ENCODED MATTER
METHOD // RML real time MATERIAL // Blue Foam MAX DEPTH // 47mm BIT // End Mill 6mm | End Mill 3mm PASSES // 2
ENCODED MATTER |
65
TEST 4.0 // CNC AND AGENT// The idea of the â&#x20AC;&#x2DC;canyonâ&#x20AC;&#x2122; triggered a significant shift in our investigation. Through the introduction of a predetermined geometry we were able to eliminate some of the unpredictability that is associated with real time milling processes. By establishing an area of intense light we were able to write agent behaviors and sensor relationships that aimed to achieve a particular form. However, this form allowed for unexpected outcomes, as the manner in which light became exposed and agent paths intersected was determined by real time processes. Because rotation and movements were based on type, the light values were not influential enough in regareds to the final geometry. The valley that formed exhibited no clear relationship to sensor readings.
Strong light readings activate type changes based on sensor readings. Z values are controlled by type.
66 | ENCODED MATTER
METHOD // RML real time MATERIAL // Blue Foam MAX DEPTH // 44mm BIT // End Mill 6mm PASSES // 2
ENCODED MATTER |
67
TEST 4.0 //
68 | ENCODED MATTER
Type: 2, 3 Depth range: -2200
Type: 1 & 3 Depth range: -375
0 -375 -625 -875 -1125 -1370 -1625 -1875 -2125 -2200
Types: 2 & 3 Input values range: 5 - 10 Depth range: -350 /- 2200
ENCODED MATTER |
69
TEST 5.0 // CNC AND AGENT// Use Introduced Geometry as Catalyst The â&#x20AC;&#x2DC;Lâ&#x20AC;&#x2122; configuration was introduced to facilitate the formation of a gentle topography. Resulting from the repetitive movements of the CNC over time, diversity of paths and z-depths accumulated at the edge of the piece due to higher sensor readings. The 6mm drill was utilized in order create a layering while still maintaining significant amount of material. The strong light readings from the curved shape at the perimeter of the material allowed for multiple Z depths to occur across a concentrated area. The rotational nature of the agent allowed for continual resurfacing and additional topological diversity.
X, Y, and Z values are based on sensor readings. The introduced geometry aims to form a slope.
70 | ENCODED MATTER
METHOD // RML real time MATERIAL // Blue Foam MAX DEPTH // 44mm BIT // End Mill 6mm PASSES // 1
ENCODED MATTER |
71
TEST 5.0 //
72 | ENCODED MATTER
Type: 1 Depth range: -1250
0 -375 -625 -875 -1125 -1370 -1625 -1875 -2125 -2200
Types: 2 & 3 Input values range: 0-2 Depth range: -350
Types: 2 & 3 Input values range: 5 - 17 Depth range: -350 /- 2200
ENCODED MATTER |
73
TEST 6.0 // CNC AND AGENT// Use Introduced Geometry as Catalyst The dual circle was introduced in an attempt to reconcile agent behavior with the practice of introducing strong geometries. Concentric circles were ideal for uniform distribution of intense light. By linking geometry to code, we were able to easily control the shift between agent behavior and geometrical formation. Unfortunately, during this simulation, agent type actions conflicted too strongly with sensor readings. The strong light readings from the curved shape at the perimeter of the material allowed for multiple Z depths to occur across a concentrated area. The rotational nature of the agent allowed for continual resurfacing and additional topological diversity.
X, Y, and Z values are based on sensor readings. The introduced geometry aims to form a slope.
74 | ENCODED MATTER
METHOD // RML real time MATERIAL // Blue Foam MAX DEPTH // 47mm BIT // End Mill 6mm PASSES // 1
ENCODED MATTER |
75
TEST 6.0 //
76 | ENCODED MATTER
Type: 1 Depth range: -2125/-2200
0 -375 -625 -875 -1125 -1370 -1625 -1875 -2125 -2200
Types: 1,2,3 Input values range: 0-1 Depth range: -350 /- 1875
Types: 2 & 3 Input values range: 4-7 Depth range: -350 /- 2200
ENCODED MATTER |
77
TEST 7.0 // CNC AND AGENT// Create an Agent Based Topography Building upon our previous study, we introduced a greater subtlety of z depths and reduced the strength of our agent behaviors. In this case, the sensor readings provided a high amount of variation because they were each linked to a specific depth. Over time this variety of z values created smooth slopes that were pocketed with deep channels carved by the circular operation as well as plateaus that resulted form sensor readings.
Agent X and Y movements are based on type but influenced by sensor. Rotation and speed change according to light intensity. Z values are determined by sensor readings.
78 | ENCODED MATTER
METHOD // RML real time MATERIAL // Blue Foam MAX DEPTH // 47mm BIT // End Mill 6mm PASSES // 1
ENCODED MATTER |
79
TEST 7.0 // Real time simulation
3000
38000
10000
45000
17000
52000
24000
59000
31000
66000
80 | ENCODED MATTER
73000
108000
80000
115000
87000
122000
94000
129000
101000
136000
ENCODED MATTER |
81
TEST 7.0 //
82 | ENCODED MATTER
Type: Drilling Circle Depth range: -2500
0 -375 -625 -875 -1125 -1370 -1625 -1875 -2125 -2200
Types: 2 & 3 Input values range: 5-17 Depth range: -375/-2200
Types: 2 & 3 Input values range: 5 - 17 Depth range: -350 /- 2200
Types: 2 & 3 Input values range: 0 - 2 Depth range: -375 /- 625
ENCODED MATTER |
83