Research On Key Technologies Of Human Motion Perception Based On Wi-Fi Radar

With the advancement and development of science and technology,humans have entered the era of the Internet of Things,and its primary purpose is to provide intelligent services centered on "people".In particular,accurate perception and understanding of human behavior is essential Technical Support.In recent years,how to accurately perceive the geographical location,trajectory,and speed of the human body in a specific environment has become the focus of attention of researchers.Traditional human motion perception systems are mostly based on computer vision,millimeter wave radar,or lidar technologies and their special special equipment,so they have the disadvantages of high cost and difficulty in large-scale application.In recent years,with the successful acquisition of channel state information by Wi-Fi devices,they also have radar-like functions and can sense changes in the surrounding electromagnetic environment.Wi-Fi radar human motion sensing technology has attracted extensive attention from researchers due to its low cost and easy deployment.However,the existing Wi-Fi radar human motion sensing technology still faces many problems that need to be resolved,mainly including:1.The signal interference in the indoor channel environment is serious,and the multipath propagation is complicated,so it is difficult to perform the channel parameter of each signal propagation path Accurate estimates.2.The accuracy of the human motion sensing method based on the characteristics of single channel parameters is low.3.The accuracy of human motion perception is greatly affected by the movement status of the human body such as position and speed.In view of the above difficult technical problems,this paper proposes a Wi-Fi radar-based human motion perception algorithm and its implementation scheme.The main research work includes the following three aspects.(1)?Research on accurate channel parameter estimation methods in complex indoor multipath environments.Firstly,for the Wi-Fi phase noise problem,the linear transformation method is used to effectively filter out the noise introduced by the hardware circuit;secondly,the signal conjugate multiplication algorithm between the antennas is used to dynamically propagate the signal containing human motion information.Path extraction;Finally,the reasons for the low accuracy of the channel parameter estimation of the classic multi-signal classification algorithm under complex multipath conditions are analyzed,and a high-dimensional joint estimation of channel parameters using space alternating generalized expected maximum algorithm is proposed.Experimental verification shows that compared with the classical multiple signal classification algorithm,the signal arrival angle estimation accuracy is improved by 59.5%,the signal flight time estimation accuracy is improved by 47.9%,and the Doppler shift estimation accuracy is improved by 39.4%.(2)?In order to further improve the accuracy of human motion perception with single-channel parameters,the human motion perception with multi-channel parameters is studied:First,this paper proposes a target signal reflection path recognition algorithm;later,the signal arrival angle,signal flight time,Doppler frequency shift jointly perceives human motion;finally,an error correction algorithm for perceptual results is proposed for possible errors in the perceptual results.Finally,more accurate multi-channel parameter human motion perception is realized.(3)?Aiming at the problem that the accuracy of human motion perception is greatly affected by the motion status of the human body such as position and speed,this paper proposes an adaptive multi-Wi-Fi link method for human motion perception:fusion of multiple Wis through an adaptive mechanism The sensing result of the Fi-link device,thereby avoiding the influence of the human body motion state on the sensing accuracy.Finally,in the actual environment,the overall scheme designed in this paper is experimentally verified.The average perception error of this scheme is 0.58m.Compared with the current best performance IndoTrack system,the accuracy of human motion perception has been improved by 37.6%.