RESEARCH ON USER BEHAVIOR PATTERN BASED ON INDOOR POSITIONING SYSTEM YUNQI WEI March - June 2018 Instructor: Li Li, Gang Yu Collaborator: Siyi Dai, Yiyang Wang Excellent Undergraduate Thesis Award
KEY WORDS: Indoor Positioning Technology(IPS); Data Acquisition; Data Mining; User Behavior Pattern As there is a trend for public buildings of being integrated and large-scaled, design problems become more and more comprehensive, and the insufficiency of tradi tional experience-based architectural design methods are increasingly evident. Indoor positioning technology (IPS) can record people's indoor activities and behaviors, to achieve quantitative analysis and provide a basis for design. In this study, UWB indoor positioning technology was adopted to monitor the using status of library reading room and collect data of user behaviors. At the same time, a series of survey is conducted to collect personal information of the users and the wireless sensor network is used to collect real-time data of building physical environment. Through the process of data mining, the study tries to explore the potential relationship between physical environmental properties, user characteristics, and user behavior patterns and a feasible advice for the improvement of the indoor design. The research process involves the complete flow from data acquisition to data mining and is a systematic attempt to apply user behavior analysis technology to the field of architectural design.
[Device Set-up & Data Acquisition]
Site: library reading-room in Southeast University Duration: May/10/18 - May/27/18
[Work Flow]
The UWB positioning device is mainly composed of tags and several base stations. The tags are used to track people’s position and communicates with the base station through the high-frequency very short pulses. In the experiment, tags are handed out to users to acquire their positioning information; the base stations are set up at a fixed position in the space, They are used to receive the very short pulses sent by tags. According to the TOA (Time of Arrival) positioning algorithm, the distance from each tag to the base stations can be measured in real time. During the data collection period, the research team’s task included: distributing the positioning tags, inviting users to complete questionnaires, visualizing data in real time, monitoring daily database conditions , cleaning invalid data, etc.
User Participation Process:
[Positioning Data Visualization & Analysis]
each tag in one day
daily selected seats
seats preference distribution
daily trace
circulation density distribution
average attendance at each hour
ATTENDANCE
hour-attendance bar graph
120 100 80 9:00AM
11:00AM
1:00PM
60 40 20
3:00PM
Receiving data packdges from positioning system through UDP
unpackaging data into tables in MySQL database
5:00PM
7:00PM
8
9 10 11 12 13 14 15 16 17 18 19 20 21 TIME
In the 18 days, we received a total of 473 valid sample data and 180,000 positioning data packages. We visualized the daily positioning data of each tag, unpackaged and cleaned the raw data received from UDP positioning sysytem, then import data into MySQL database for storage, so it can be called at any time we need during data analysis. We overlaid 18 days results together, then got the seats preferences distribution, circulation density distribution, and each hour attendance respectively.
[Physical Environment&Energy Conservation] We arranged 18 wireless sensors evenly to collect physical environment data of the reading room every two minutes. Then we did the mapping concerning average temperature, humidity, and illumination per hour. Through cross-analyzing the environment data and students’ attendance we can find some regular rules and accordingly propose reasonable suggestions for energy saving plan(for example, there are few people in the room before 9, so there is no need turning on the air condition). What is more, if data can be processed in real time, it will be possible for the whole equipment system such as lighting and air system, to give real-time feedback.
Please fill in your TAG#: ______09________ (Timestamp: 2018-05-19 10:31:00)
User Information
Positioning Data
[User Information & Positioning Data]
Tag ID: 09
We linked users’ information that we acquired from the electronic questionnaires and their positioning Information (such as selected seats, residence time, and activity traces ) one-to-one through TagID and Timestamp. By cross-analysis of users’ data and positioning data, we are able to reveal those unsatisfied demands and current design issues and have a better understanding of human behavior.
Timestamp: 2018-05-19 10:31:00
Q: What is your main consideration when you are selecting seats?
Q: Do you come alone or with friends?
Q: What are you coming here for?
Q: How often do you come here?
[Summary & Future Prospect] On the one hand, by analyzing the data acquired from Indoor Positioning System, we can have a specific understanding of human behavior, which can provide the basis for design for space designers at the early stage. On the other hand, Indoor Positioning System helps us get real-time information about people’s activities in a particular space. Together with Building Automatic System (BAS), space itself gets the intelligence to give real-time feedback and always keeps itself in best quality. Overall, Indoor Positioning System fills the gap where the Global Positioning System (GPS) loses its effect. It will play an increasingly indispensable role especially in large-scale buildings where indoor space is complex such as medical building, mega-malls, and large exhibition hall.
Users real-time feedback
IPS design basis
Designers
[Re-layout Based On Data Analysis]