An ESPRIT-like algorithm for angle estimation in bistatic MIMO radar

This paper proposes a novel ESPRIT(Estimation of Signal Parameters via Rotational Invariance Techniques)-like algorithm for DOD (Direction of Departure) and DOA (Direction of Arrival) estimation in bistatic MIMO (Multiple Input Multiple Output) radar in order to reduce computational burden of conventional methods based on the fourth-order cumulant. Firstly, the fourth-order cumulant matrix of received data is computed which extends the aperture and restrains Gaussian colored noise. Secondly, considering of the high dimensions of virtual steering vector, an exchange matrix and a reduced-dimensional matrix are employed to transform the data matrix into a low dimensional space with the structure of conjugate symmetry. Thirdly, the real signal space and noise space are got from the complex covariance of virtual array. Then DOD and DOA are estimated, based on the rotational invariance techniques in the real number field. The proposed algorithm reduces the data dimensions from M²N² to (2M −1)(2N −1), then estimates angles in the real number field, which can greatly avoid high computational cost. It also extends antenna aperture, and restrains the influence of the colored noise. The numerical simulation results demonstrate the efficiency of the proposed method.