With the development of the Internet of Things(IoT) and the popularization of commercial WiFi,researchers have begun to use commercial WiFi for human activity recog-nition in the past decade.However,cross-scene activity recog-nition is still difficult due to the different distribution of sam-ples in different scenes.To solve this problem,we try to build a cross-scene activity recognition system based on commer-cial WiFi.Firstly,we use commercial WiFi devices to collect channel state information (CSI) data and use the Bi-directional long short-term memory (BiLSTM) network to train the activity recognition model.Then,we use the transfer learning mecha-nism to transfer the model to fit another scene.Finally,we con-duct experiments to evaluate the performance of our system,and the experimental results verify the accuracy and robustness of our proposed system.For the source scene,the accuracy of the model trained from scratch can achieve over 90％.After transfer learning,the accuracy of cross-scene activity recogni-tion in the target scene can still reach 90％.

Yuanrun FANG;Fu XIAO;Biyun SHENG;Letian SHA;Lijuan SUN

School of Computer Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China

计算机科学前沿

[1]Yuanrun FANG;Fu XIAO;Biyun SHENG;Letian SHA;Lijuan SUN-.Cross-scene passive human activity recognition using commodity WiFi)[J].计算机科学前沿,2022(01):83-93

Cross,passive,human,activity,recognition,using,commodity,WiFi,With,development,Internet,Things,IoT,popularization,commercial,researchers,have,begun,use,past,decade,However,cross,still,difficult,due,different,distribution,sam,ples,scenes,To,solve,this,problem,try,build,system,Firstly,devices,collect,channel,state,information,CSI,data,directional,long,short,term,memory,BiLSTM,network,model,Then,transfer,learning,mecha,nism,fit,another,Finally,con,duct,experiments,evaluate,performance,experimental,results,verify,accuracy,robustness,proposed,For,source,trained,from,scratch,can,achieve,over,After,target,reach

0.517286