The Study On Array Diagnosis And Corrupted Data Recovery In Bistatic MIMO Radar

A Multiple-Input Multiple-Output(MIMO)radar achieves better target detection and parameter estimation performance compared with the traditional phased radar,which has been widely researched in recent years.As the transmit and receive array of MIMO radar are consisting of numerical elements,it is highly probability that the faulty element appear in the array with the impact of harsh environment.Both faulty element in the transmit and receive array lead to the corrupted data of virtual array formed by the matched filters.In this case,angle estimation performance of classical methods may degrade seriously.To improve the array performance against to the faulty elements and corrupted data,array diagnosis method and data recovery method for angle estimation are studied in this thesis.The main contributions and innovations can be concluded as following:(1)These methods require setting external probes for obtaining a large number of measurements,and,consequently,are time-consuming in the case of large arrays due to a long measurement time and a long detection time.Aiming at extra probes are required to collect numerical data,an array diagnosis method without extra probes is proposed based on the similarity between received signal.The abnormal data collected by the virtual array has barely relation with the abnormal data.Following this point,the cross-correlation method,which measures the dissimilarity between the outputs of the perfect virtual sensor and the invalid virtual element,is used to locate the impaired element in both transmit array and receive array.(2)The array diagnosis for bistatic MIMO radar with the mutual couple present is studied.Mutual coupling error is first taken into account for the signal model with the impaired element in the transmit array and receive array.First,the signal subspace of covariance of virtual array is acquired by the eigenvalue decomposition.Second,according to the data distribution characteristics,the data points in the subspace are automotive clustered by the density peaks cluster algorithm.Finally,the anomaly cluster is distinguished and both the impaired element in the transmit array and receive array are detected based on the cluster result.(3)To recover the corrupted data in covariance matrix,a method is proposed to recovery the intact covariance matrix of virtual array from the corrupted one.By designing a 4-fold Hankel from the corrupted covariance matrix of virtual array,the zero rows/columns caused by the impaired element are eliminated and the low rank property is also kept which means the missing element in the designed matrix can be recovered by matrix completion.Following the recovered 4-fold Hankel matrix,the corrupted data in the covariance matrix is then fulfilled by the inverse transformation of completed 4-fold Hankel matrix.The rotation invariance of signal subspace in completed covariance matrix is following restored and the better angle estimation performance of ESPRIT is achieved.