Research On Multidimention Angle Estimation In MIMO Radar With L-shaped Array

L-shaped array MIMO radar can estimate azimuth and elevation angle of targets, which overcomes the disadvantage of estimating only one-dimension angle in uniform linear array MIMO radar. The computational complexity is huge of two-dimension angle estimation algorithms directly applied with traditional one-dimension angle estimation algorithms, which serious influence the real-time performance of radar system. Inspired by the ideas and theories of traditional angle estimation algorithms and the equivalent virtual array of L-shaped array MIMO radar, some DOA estimation methods with low-complexity and ones under single snapshot are proposed in this dissertation.Firstly, the equivalent virtual array structure, the DOF, the maximum number of resolvable targets and the Cramer-Rao bound of L-shaped co-location transceiver array MIMO radar and L-shaped division transceiver array MIMO radar are analyzed according to the signal models of the two L-shaped array MIMO radar. Then the typical DOA estimation algorithm in array signal processing theory is summarized through simulations and comparisons.Secondly, in order to decrease the computational complexity, the reduced-dimension MUSIC algorithm is proposed to reduce the two-dimensional spatial spectrum searching into one-dimensional spectral searching based on the theories of MUSIC algorithm and the feature of the equivalent virtual array in L-shaped co-location transceiver array MIMO radar. Then the reduced-dimension PM-MUSIC algorithm is proposed to avoid the covariance matrix decomposition, which is conductive to project implementation. Inspired by the ideas and theories of ESPRIT algorithm and PM algorithm, the PM-ESPRIT algorithm is put forward,which could reduce computational complexity to a certain extent. At last, the unitary ESPRIT algorithm is proposed based on the unitary transformation and the reconstructed signal, throw which the covariance matrix of reconstructed sigal is a centro Hermitian matrix. Then the complex-value covariance matrix can be turned into a real-value field by unitary transformation. Thus, the computational complexity is lower and the precision is higher.Thirdly, a reduced-dimension matrix is designed to eliminate the redundant array elements in target echo signal,which can minish the dimension of the target echo signal in L-shaped division transceiver array MIMO radar. Then the reduced-dimension MUSIC algorithm is proposed to reduce the two-dimensional spatial spectrum searching into one-dimensional spectral searching, which can highly decrease the computational complexity.Then based on the ideas and theories of ESPRIT algorithm,the reduced-dimension ESPRIT method is proposed to reduce the computational cost.Finally, considering of the problem of two-dimension DO A estimation under single snapshot, the 2-D unitary matrix pencil method is proposed based on the equivalent virtual rectangular array of the L-shaped co-location transceiver array MIMO radar and the L-shaped division transceiver array MIMO radar. Becouse of the unitary transformation, the method not only has lower computational complexity,but also has a higher precision under single snapshot.