Research On Target Detection And DOD-DOA Estimation For MIMO Radar Based On Riemannian Manifold

Multiple-input multiple-output(MIMO)radar can use multiple antennas at the transmitter to simultaneously transmit different waveforms and use multiple antennas at the receiver to simultaneously receive echo signals.Therefore,compared with traditional radar systems,its target detection and estimation performance is superior.In MIMO radar system,target detection and angle estimation are the basis for subsequent operations such as target positioning and recognition,which have important research significance.In order to improve the performance of detection and estimation,scholars at home and abroad have studied it deeply and put forward many ideas and algorithms,but there are still some defects to be overcome.First,when the signal-to-noise ratio(SNR)is relatively low,the performance of traditional target detection and estimation is obviously reduced.Second,sample covariance matrix can no longer replace statistical covariance matrix when the number of samples is small or even single-snapshot.In addition,the traditional methods are processed in Euclidean space,which cannot make full use of the information about the signal matrices and the noise matrices,and it will affect the performance of the algorithm.Therefore,based on the characteristic that Hermitian positive definite matrices will form Riemannian manifolds in space,using Riemannian manifolds and burg recursion method as tools,we have conducted in-depth research on target detection and DOA and DOD estimation of monostatic MIMO radar in order to improve the detection and estimation performance.The creative works of the thesis are as follows:Aiming at the problem of poor detection performance at low SNR,a novel target detection method based on Riemannian manifold is proposed for monostatic MIMO radar.According to three different definition of Riemannian means,the estimate of the noise covariance matrices is respectively calculated in this method.Then,theRiemannian distance between the Riemannian mean and the received signal covariance matrix is taken as the detection statistic.The decision threshold is derived based on the false alarm probability and the statistical distribution of noise.Simulation results show that the algorithm can effectively improve target detection performance.In view of the fact that there is single-snapshot in the actual environment,a target detection method for monostatic MIMO radar based on burg recursion and Riemannian manifold is proposed.In the method,the single-snapshot Toeplitz-Hermitian positive definite(THPD)covariance matrix of received signals is generated through the burg recursion algorithm,and the Riemannian mean of the noise covariance matrices is calculated.The Riemannian distance between them is used as the detection statistic.The simulation results indicate that compared with the traditional Euclidean distance-based method,the proposed method can significantly improve the target detection performance under low SNR and single-snapshot.Since the target DOA and DOD estimation algorithm based on the traditional MUSIC have some limitations,it is unable to estimate the target effectively when the SNR is relatively low and the number of snapshots is small.Therefore,a target DOA and DOD estimation method of monostatic MIMO radar based on burg recursion and Riemannian manifold is proposed.This method first obtains the Riemannian mean,which is calculated by the statistical covariance matrices of the received signal,then uses the transmit-receive joint steering vector to generate the steering covariance matrix.The Riemannian distance between them can be used to construct a spectral peak search function,so as to achieve the target DOA and DOD estimation of MIMO radar with low SNR and few snapshots.