Understanding and Modeling of WiFi Signal Based Human Activity Recognition Wei Wangt,Alex X.Liutt,Muhammad Shahzadf,Kang Lingt,Sanglu Lut t Nanjing University,Michigan State University September 8,2015 日,+司1三13卡三分Q0 1/24
Understanding and Modeling of WiFi Signal Based Human Activity Recognition Wei Wang† , Alex X. Liu†‡, Muhammad Shahzad‡ ,Kang Ling† , Sanglu Lu† †Nanjing University, ‡Michigan State University September 8, 2015 1/24
Motivation Modeling Design Experiments Conclusions Motivation WiFi signals are available almost everywhere and they are able to monitor surrounding activities. en5 立)Q0 224
Motivation Modeling Design Experiments Conclusions Motivation • WiFi signals are available almost everywhere and they are able to monitor surrounding activities. 2/24
Motivation Modeling Design Experiments Conclusions Problem Statment WiFi based Activity Recognition Using commercial WiFi devices to recognize human activities. Wireless router Laptop Advantages √Work in dark Better coverage Less intrusive to user privacy No need to wear sensors 日,+司1三13卡三分Q0 3/24
Motivation Modeling Design Experiments Conclusions Problem Statment WiFi based Activity Recognition • Using commercial WiFi devices to recognize human activities. Wireless router Laptop Wireless signal reflection Advantages X Work in dark X Better coverage X Less intrusive to user privacy X No need to wear sensors 3/24
Motivation Modeling Design Experiments Conclusions Problem Statment WiFi based Activity Recognition Using commercial WiFi devices to recognize human activities. Wireless router Laptop Advantages Work in dark Better coverage Less intrusive to user privacy No need to wear sensors 至aC 3/24
Motivation Modeling Design Experiments Conclusions Problem Statment WiFi based Activity Recognition • Using commercial WiFi devices to recognize human activities. Wireless router Laptop Wireless signal reflection Advantages X Work in dark X Better coverage X Less intrusive to user privacy X No need to wear sensors 3/24
Motivation Modeling Design Experiments Conclusions Limitations of Prior Arts Limitations of Prior Arts:no model(signal,activity) So,have to rely on statistical characteristics of WiFi signals So,sensitive to environmental changes signals Our Approach:model(signal,speed)+model(signal.activity) Robust to environmental changes High recognition accuracy 日,+司1三13卡三分Q0 424
Motivation Modeling Design Experiments Conclusions Limitations of Prior Arts Limitations of Prior Arts: no model (signal, activity) • So, have to rely on statistical characteristics of WiFi signals • So, sensitive to environmental changes signals Our Approach: model(signal, speed)+model(signal, activity) • Robust to environmental changes • High recognition accuracy 11 11.5 12 12.5 60 65 70 75 CSI Time (seconds) 4/24
Motivation Modeling Design Experiments Conclusions Limitations of Prior Arts Limitations of Prior Arts:no model(signal,activity) So,have to rely on statistical characteristics of WiFi signals So,sensitive to environmental changes signals Our Approach:model(signal,speed)+model(signal,activity) Robust to environmental changes High recognition accuracy h rIW 115 Time(seconds) Dac 4/24
Motivation Modeling Design Experiments Conclusions Limitations of Prior Arts Limitations of Prior Arts: no model (signal, activity) • So, have to rely on statistical characteristics of WiFi signals • So, sensitive to environmental changes signals Our Approach: model(signal, speed)+model(signal, activity) • Robust to environmental changes • High recognition accuracy 11 11.5 12 12.5 60 65 70 75 CSI Time (seconds) 4/24
Motivation Modeling Design Experiments Conclusions Understanding Multipath Key observations Sender Multipaths contain both static component and dynamic com- LoS path ponent Reflected by wall d,(0) Each path has different phase Phases determine the ampli- Reflected by body tude of the combined signal Wall Receiver 日,+司1三13非三分Q0 5/24
Motivation Modeling Design Experiments Conclusions Understanding Multipath Key observations • Multipaths contain both static component and dynamic component • Each path has different phase • Phases determine the amplitude of the combined signal Sender Wall Receiver Reflected by body Reflected by wall LoS path dk (0) dk (0) 5/24
Motivation Modeling Design Experiments:Conclusions Understanding Multipath 00 Sender Q LoS path Dynamic Q Static Component Reflected by component wall d(oj Reflected by Combined body Wall Receiver 口,01三13卡至QC 6/24
Motivation Modeling Design Experiments Conclusions Understanding Multipath Sender Wall Receiver Reflected by body Reflected by wall LoS path dk (0) dk (0) I Q Combined Static component Dynamic Component 6/24
Motivation Modeling Design Experiments:Conclusions Understanding Multipath Sender Q Dynamic Component d(t) LoS path Static 】 component Reflected by wall Combined Reflected by body Wall Receiver 日,+司1三13卡三分Q0 6/24
Motivation Modeling Design Experiments Conclusions Understanding Multipath Sender Receiver dk (t) Wall Reflected by body Reflected by wall LoS path LoS path I Q Combined Static component Dynamic Component 6/24
Motivation Modeling Design Experiments. Conclusions Understanding Multipath Sender Q+ Combined Dynamic Component d.) LoS path Static Reflected by component wall Reflected by body Wall Receiver 日,+司1三13卡三分Q0 6/24
Motivation Modeling Design Experiments Conclusions Understanding Multipath Sender Receiver dk (t) Wall Reflected by body Reflected by wall LoS path I Q Combined Static component Dynamic Component 6/24