Research On Passive Indoor Human Detection And Localization Technology Based On CSI

With the development of network technology and the widespread popularity of Wi-Fi infrastructure,mobile smart devices and wireless networks have influence all aspects of human production and life,and location-based service(LBS)has gradually become an indispensable part in people's lives.Localization is the most important part of LBS.As cities develop,the living environment of modern people has gradually shifted from outdoor to indoor.Although the satellitebased outdoor localization technology has been relatively mature and has been widely used.However,the severe decline in satellite signals caused by reinforced concrete makes it difficult to navigate and localize indoor.Therefore,it is of great significance and value to seek new signals to be applied to indoor human detection and localization technology.In recent years,Wi-Fi-based indoor localization has received extensive attention in the academic community.However,most Wi-Fi-based human detection and indoor localization systems have the disadvantages of being greatly affected by the environment,low accuracy,complex models,and requiring users to carry specific equipment,which limit the universality of these systems.Therefore,it is particularly important to design a universal and reliable passive indoor human detection and localization system.In human detection,this thesis proposes a passive indoor human detection system based on WiFi.This system is based on channel state information(CSI),CSI feature fingerprints are generated after date preprocessing.The system uses BP neural network to determine whether the specific state of the indoor scenario is no target,a static target or a dynamic target.In order to improve the accuracy of human detection,this research makes full use of multiple input multiple output(MIMO)information,A multi-antenna voting scheme is adopted in the simulation experiment which comprehensively considers the detection results of different antenna pairs.The experimental results show that the passive indoor human detection system proposed in this research has a high detection accuracy.The accuracy can reach more than 99% when there is a static target or no target indoor.The accuracy of scenario with a dynamic target indoor can reach more than 98%.The system also has a low false alarm rate.In indoor localization,this thesis proposes a depthwise separable convolution based passive indoor localization system(DSCP)using Wi-Fi CSI.DSCP is a fingerprint-based localization system,which includes an offline training phase and an online localization phase.In the offline training phase,the indoor scenario is first divided into different areas to set training locations for collecting CSI.Then amplitude differences of these CSI subcarriers are extracted for constructing location fingerprints,thereby training the convolutional neural network(CNN).In the online localization phase,CSI data is first collected at the target locations,then the location fingerprint is extracted and finally fed to the trained network to obtain the predicted location.The experimental results show that DSCP has a short training time and a low localization delay.DSCP achieves a high localization accuracy,upper than 97%,and a small median localization distance error of 0.69 m in typical indoor scenarios.