Research In Angle Estimation For MIMO Radar Based On Compressed Sensing

As a new radar system, multiple input multiple output(MIMO) radar has a guiding significance for better understanding of the traditional radar and research of the new radar system because of its superiority of system. Compressed Sensing(CS) theory, as a new signal processing technology, has been applied in radar signal processing by many researchers. In this thesis, the Compressed Sensing theory is combined with the angle estimate of MIMO radar, the main works are summarized as follows:1. Research on spatial spectrum estimation for MIMO radar. The traditional algorithms including Capon, MUSIC and ESPRIT, are promoted to direction of department(DOD) and direction of arrival(DOA) estimation for Bistatic MIMO Radar. In addition, to solve the problem of angle estimation for MIMO radar in the presence of the colored noise, a new method based on improved fourth-order cumulants(FOC) for joint DOD and DOA estimation was proposed. The redundant information of the echoed signal was removed to reduce its dimension, while the effective aperture of the virtual array kept unchanged. Finally, angle estimation was realized according to MUSIC-like based on FOC. The proposed method can suppress the effect of Gaussian colored noise, and reduce the computational complexity by reducing the dimension of the FOC matrix.2. CS theory is applied in angle estimate of MIMO radar, and a new method is proposed for the joint DOD and DOA estimation of bistatic MIMO radar based on sparse recovery. Firstly, a redundant dictionary is built based on the two-dimension angle scene. Then, effective eigenvectors obtained from the covariance matrix of array received signals are sparse denoted in the redundant dictionary, and then a low-dimensional sparse linear model is constructed. Finally, angle estimation will be obtained by the sparse recovery algorithm. The proposed algorithm reduces the computational complexity of directly reconstructing the original signals and performs well even under the low SNR and low snapshots.3. Based on high-order cumulants combined with CS theory, a new method for joint DOD and DOA estimation based on high-order cumulants and sparse representation was proposed in the presence of colored noise. The Gaussian colored noise could be suppressed by the FOC which is insensitive to Gaussian noise. The signal subspace is derived by eigenvalue decomposition of the FOC matrix, and the eigenvectors forming the signal subspace are sparsely denoted in the appropriate redundant dictionary. Finally, joint DOD and DOA estimation was realized according to solving the sparse coefficients by sparse reconstruction algorithm. Simulation results verify that the proposed method suppressed the effect of the Gaussian colored noise, and performs steady even under the low snapshots.