The Study On Array Diagnosis And Angle Estimation In Bistatic MIMO Radar Under Array Antenna Failures

Multiple-input multiple-output(MIMO)radar system has more and more antennas in the enlarged transmit array and receive array.However,it is highly probability that the failure elements appear in bistatic MIMO radar system due to harsh environment and hardware aging in practical application.In this case,the traditional angle estimation algorithms have angle estimation performance degradation or even fail to work.In order to solve this problem,a signal-processing scheme for bistatic MIMO radar with impaired array is investigated in this thesis,including array diagnosis as well as angle estimation.The main contribution can be concluded in following aspects:(1)The current researches typically attempt to diagnose array with the far-field or nearfiled pattern measurements collected by waveguide probe.However,the use of waveguide probe increases hardware costs,and interferes with normal array operation.Therefore,the methods for array diagnosis without waveguide probes should be developed.The modulus of the outputs of virtual array in bistatic MIMO radar can be considered as a grayscale image.Since the received data of normal element contains the target signal and noise signal,the normal element has higher image edge strength compared with damaged element.Therefore,the array diagnosis can be achieved taking advantage of the difference of image edge strength between normal elements and damaged elements.In order to improve the real-time performance of array diagnosis,an image entropy based array diagnosis method is proposed.The positions of fault elements are located based on the characteristic,which is that the image entropy of the received data of the normal antenna is larger than that of the received data of the damaged antenna.(2)Due to the fault antenna elements in bistatic MIMO radar,the successive data missing occurs along the row and column directions of the covariance matrix,which results in the degradation of angle estimation performance.A covariance matrix reconstruction based angle estimation method is proposed.According to the block Toeplitz property of covariance matrix in bistatic MIMO radar,the missing elements in the sub-block matrix of the covariance matrix caused by the damaged antennas can be reconstructed from the intact sub-block elements.Then a complete covariance matrix is obtained to improve the angle estimation performance at low SNR for bistatic MIMO radar with damaged antennas.(3)In order to improve the angle estimation performance for bistatic MIMO radar under array antenna failures,a signal subspace reconstruction based angle estimation method is proposed.Firstly,the signal sub-matrices are obtained by dividing the signal subspace according to the number of transmit elements.A low-rank block-Hankel matrix is constructed based on the relevance between the signal sub-matrices.The matrix completion(MC)can be then be applied to accurately recover the complete signal subspace matrix.Therefore,the rotation invariance structure in the recovered signal subspace is restored,which facilitates use of the standard ESPRIT algorithm for angle estimation in bistatic MIMO radar under array antenna failures.The angle estimation accuracy of the proposed method is very close to that of ESPRIT algorithm implemented for the normal array case.