Signal Processing Methods And System Design Of Through-Wall Human Motion Detection

Over the past decade,the technology of through-wall human motion detection mainly uses the ultra wideband signal,which usually adopts GHz bandwidth,high power in order to distinguish the objects behind the wall,and radar system has generally a large size.If not use the ultra wideband signal,instead of making use of the signal of Industrial Scientific Medical(ISM)frequency band to realize through-wall human motion detection system that has the small bandwidth,low power,compact size and low cost will have a significant influence on all aspects of social public safety and people's lives.In this paper,techniques of through-wall human motion detection are deeply researched,and a through-wall human motion detection system based on ISM signal is designed and implemented.The main work of this paper includes:1.The through-wall human motion detection system is designed.Firstly,the mechanism of through-wall human motion detection under strong background noise interference is studied.Then leveraging IEEE 802.1 In signal to communicate and pre-coding to eliminate strong interference signals reflected from walls and other static objects.Thus the weak signal of moving human behind the wall can be extracted.Finally,utilizing Software Defined Radio equipment and platform to achieve signal transmission and reception as well as the infer-ence suppression of background noises.By means of analyzing the effect of interference suppression,it is concluded that the suppression effect has an upper limit.2.Off-line signal processing methods for identifying moving human through the wall are re-searched.Via processing and analyzing receiving signals in time domain,frequency domain and joint time-frequency domain,feature attributes can be extracted for detection.Then a method of through-wall moving human detection based on the threshold is proposed.In or-der to be applied to different detection environments,a method is devised to detect moving human through wall based on cluster and classification by introducing the idea of machine learning.3.Off-line signal processing methods for through-wall human motion orientation are explored.By obtaining channel state information from the received signal and then the Doppler fre-quency shift,it can judge the moving direction relative to the receiver.Moreover,the moving velocity of human can be estimated roughly by the frequency shift.In addition,using the algorithm of dynamic time warping can able to detect and make a distinction between hori-zontal and vertical motion along the wall,which are two motion orientations.4.Real-time signal processing methods for through-wall human detection are researched,and the results display interface is excogitated.Firstly,Via modifying off-line methods and com-ing up with lightweight means can give motion mode information in a much shorter time range.Then the results display interface and off-line analysis tool are designed.Finally,the detection accuracy is analyzed and discussed.