Target Location Algorithm For Through-wall Radar Based On High Resolution Range Profile

The use of electromagnetic waves to detect targets after non-transparent obstacles has become an effective and practical method,which has been widely used in military and civilian fields.Through-Wall Radar(TWR)is now widely used in indoor detection missions.It can be used to detect and locate enemies behind walls.It can also detect the breathing and heartbeat of the survivors in disasters,so the search and rescue work can be carried out.To achieve accurate through-wall detection,in addition to designing a reliable radar system,it is also necessary to solve the key problems encountered in wall penetration detection: non-linearity;morbidity;real-time;wall effect.In this paper,FDTD is used to simulate the propagation characteristics of electromagnetic waves through the wall,and the four key problems in the detection of through-wall are studied.The research results are as follows:Firstly,the through-wall propagation characteristics of ultra-wideband(UWB)signals are studied.It is verified that electromagnetic waves have experienced time delay,signal attenuation and waveform distortion in the process of penetrating the wall.Secondly,Compressive Sensing(CS)is used to extract the High Resolution Range Profile(HRRP)feature of the target behind the wall.Reconstruction algorithm is a difficult point in CS theory implementation.This paper uses Orthogonal Matching Tracking(OMP)algorithm,Bayesian Compressed Sensing(BCS)algorithm and Multi-Task Bayesian Compressed Sensing(MTBCS)algorithm are used to extract HRRP features of the target behind the wall.The simulation results show that the target HRRP extracted by the OMP algorithm contains the least clutter.Secondly,this paper studies the influence of wall information and target information on target HRRP feature extraction.The simulation results show that the target HRRP can be used to invert the target position and size.Thirdly,a real-time method based on Support Vector Machine(SVM)realizes the problem of through-wall positioning.This method does not need to estimate the role of the wall,and the nonlinearity and ill-posedness in the wall-passing problem can also be solved.Since the problem of through-wall positioning can be transformed into a mapping relationship between the target HRRP and the target information,the nonlinearity in the wall-passing problem and the through-wall propagation effects(time delay,signal attenuation,and waveform distortion)are included in this mapping relationship.In the meantime,the hyperparameter regularization in the SVM can solve the ill-posedness in the wall penetration problem.Simulation results show that the method is effective and robust.Fourthly,the wall echo suppression is a key problem that must be solved in the wall positioning.It is especially difficult to suppress the wall echo when the target is stationary.The Singular Value Decomposition(SVD)method based on subspace projection can effectively eliminate the wall echo,but it is difficult to extract the subspace features corresponding to the weak target.In contrast,the Independent Component Analysis(ICA)method based on statistical independence and non-Gaussian can better extract the independent components representing weak target signals.Combined with the SVD method and the ICA method,this paper proposes a wall echo suppression method.The simulation results show that the proposed method can effectively suppress the wall echo for the smooth or rough uniform medium wall.