Analysis On Living-Lifeless Human Body Detection Methods Based On MIMO Radar

Living-lifeless human body detection is a technique to distinguish between the living human body and the lifeless human body,which has been widely used in military conflict,disaster rescue,security monitoring,and other fields.Radar-based human target detection has the advantages of strong penetration and resistance to interference from temperature and noise,etc.Living-lifeless human body detection is mainly to detect the human movement caused by vital signs such as respiration and heartbeat,which are generally weak.Therefore,living-lifeless human body detection belongs to the category of weak target detection,which has always been a challenging problem in radar signal processing.Multiple-input multiple-output(MIMO)radar can improve target detection performance.Hence,this thesis works on the analysis of living-lifeless human body detection methods based on MIMO radar.The main contributions and innovations are as follows.First,this thesis proposes mathematical models for living-lifeless human body detection based on MIMO radar.Different from the traditional target echo signals with time-invariant delay and phase,this thesis develops a MIMO radar echo signal model with time-varying delay and phase for the living human target based on the physiological characteristics of human respiration.The binary composite hypothesis test problem of the living-lifeless human body is formulated for the scenario where chest fluctuation amplitude caused by respiration,respiratory frequency,and target reflection coefficient are deterministically unknown.Due to the weak human respiration motion,the degree of difference between the null hypothesis and the alternative hypothesis may be too small to distinguish between them.In order to describe the different degrees of difference between the two hypotheses,this thesis introduces the concepts of far hypotheses and close hypotheses.Then,in view of the fact that there is no uniformly most powerful test(UMPT)for the composite hypothesis test problem of the living-lifeless human body,the generalized likelihood ratio test(GLRT)for the far hypotheses and the generalized locally most powerful test(GLMPT)for the close hypotheses are proposed respectively.The closed-form expression of the detection probability of the GLMPT detector with the known respiration frequency is derived.Meanwhile,the performance of the GLRT detector under the far hypothesis and the GLMPT detector under the close hypothesis are analyzed respectively.Finally,the effects of carrier frequency,signal duration,and the number of receiver antennas on the far-close property of the two hypotheses,as well as the performance of the GLMPT and GLRT detectors under both far and close hypotheses are analyzed.Furthermore,this thesis analyzes the effect of non-Gaussian clutter plus noise on the detection performance.All of these analyses provide guidance for the optimal design of radar systems.