Human Target Detection And Localization Based On Multi-Channel Millimeter Wave Radar

Multi-channel millimeter-wave radar can detect the related parameters of human target,such as target number,range,speed and other relevant information in the radar detection range,which have a wide range of applications in different fields,like smart home,intelligent transportation,unmanned monitoring and so on.This thesis focuses on issues such as human micro-motion parameter estimation,walking human detection,and walking human localization in real time.The specific work is as follows:1.Aiming at the problem of human micro-motion parameter estimation,an echo model of micro-motion human echo was established,and a micro-motion parameter estimation algorithm based on multi-channel cross-correlation entropy was proposed.The algorithm first performs the autocorrelation entropy calculation on the echo signals of a single channel;then,after performing the two-two cross-correlation processing on the above-mentioned autocorrelation entropy matrix,it performs arithmetic fusion to obtain the multi-channel cross-correlation entropy matrix;and finally,the cross-correlation entropy.The matrix is imaged and the human fretting parameters are estimated in the image domain.Compared with the traditional Fourier analysis algorithm and autocorrelation entropy algorithm,this algorithm can effectively improve the signal-to-noise ratio of micro-motion target imaging.The simulation results verify the performance of the algorithm.2.For human walking detection problems,a model of moving human echo is established,and a parameter estimation method based on the range-Doppler 2D image domain is improved.Through a targeted distance-velocity decoupling strategy,such as monotone frequency-distance coupling for slow walking human target and triangular frequency-modulation solution for distance-velocity coupling for rapid walking,distance and velocity parameter estimation accuracy are improved.The theoretical analysis and simulation give the verification;Provide the performance analysis of target detection algorithm based on CA-CFAR and the extended target coalesces algorithm.Simulation and measured data validation are performed.3.Aiming at the real-time positioning problem of walking around the human body,the human walking target positioning algorithm based on phase comparison was studied.The method extracts the target phase information of different channels,calculates the phase information difference between the ring channels,and converts it into the angle information of the coordinates of the target.In combination with the target distance measurement method,the coordinate information of the target is calculated to achieve the target positioning.Compared with time difference positioning method and DOA estimation algorithm,the method has a simple principle and high positioning accuracy,can realize real-time positioning,is suitable for small antenna arrays,and is easy to implement in engineering.This thesis focuses on the problems in the practical application scenario of multi-channel millimeter-wave radar,combines simulation and measured experiments,and conducts related research on human micro-motion parameter estimation,human walking detection,and real-time positioning of the human body,etc.,and is related to the follow-up human body tracking technology.Research laid a solid foundation.