Research On Angle Estimation And Array Calibration In MIMO Radar System

Multi-input multi-output (MIMO) radars which utilize multiple antennas to simultaneously transmitdiverse waveforms and receive reflected signals in similar ways, have the potential advantages overconventional phased-array radar on performance of targets detecting and parameter estimating.Direction of arrival (DOA) estimation is a key technology in MIMO radar. In practical MIMO radarsystem, there may exsit some cases like the array is abiratry, the number of snapshots is small andeven the antennas in transmit/receive array have gain-phase errors. For these possible cases in praticalapplication, we investigate approaches of angle estimation and array Calibration in MIMO RadarSystems. The main work in this paper is summarized as follows:(1) A joint two-dimensional direction of departure (2D-DOD) and two-dimensional direction ofarrival (2D-DOA) estimation algorithm in bistatic MIMO radar with arbitrary arrays is proposed. Thealgorithm is to extend the propagator method (PM) for angle estimation which does not require peaksearching and eigenvalue decomposition of received signal covariance matrix. Therefore, it has lowcomputational complexity. And it can achieve automatically pairing four-dimensional angles.Furthermore, the proposed algorithm has much better angle estimation performance than interpolatedestimation method of signal parameters via rotational invariance techniques (ESPRIT), and has veryclose angle estimation performance to ESPRIT-like algorithm which has higher computational costthan the proposed algorithm.(2) We investigate the topic of coherent angle estimation in MIMO radar, and propose alow-complexity algorithm based on subarray spatial smoothing for coherent DOA estimation inmonostatic MIMO radar and a scheme for coherent angle estimation by exploiting a parallel profilewith linear dependencies (PARALIND) decomposition in bistatic MIMO radar. The former has lowcomplexity and better DOA estimation performance than some conventional coherent angleestimation algorithms. And the latter not only realizes precise coherent angle estimation,but alsoacquires correlated coefficient matrix of sources.(3) We make further efforts to discuss the coherent angle estimation in monostatic MIMO radarusing a single pulse. A reduced dimentional (RD) ESPRIT and a trilienear decomposition basedcoherent angle estimation algorithm are proposed on this issue. These two algorithms use the receiveddata to construct a set of Toeplitz matrices through which RD-ESPRIT or trilienear decompositionwas utilized to estimate DOA. The two proposed algorithms are all effective for coheren angleestimation using a single pulse and have considerable performance.(4) A practical issue of DOD and DOA estimation for MIMO radar with array gain-phase errors is considered. In terms of this issue, a PM-like algorithm for joint angle and array gain-phase errorsestimation is proposed which not only reaches automatically paired estimates of the angles andgain-phase errors but also has much better gain-phase errors estimation performance than theESPRIT-like algorithm which has higher computational cost than the proposed algorithm.Furthermore, the proposed algorithm has very close angle estimation performance to ESPRIT-likealgorithm.