A Tensor Method of Angle Estimation for Bistatic MIMO Radar in the Presence of Spatially Colored Noise and Strongly Correlated Targets

This paper is devoted to the angle estimation problem for bistatic multiple-input multiple-output (MIMO) radar under strongly correlated targets and spatially colored noise scenes. Firstly, the cross-covariance matrix is established by using the independent noise vector after matched filtering in time domain, and the components of spatial colored noise are eliminated. Then, the cross-covariance matrix is rearranged into a third-order tensor, and the reconstructed tensor can be obtained by concatenating two third-order tensors. Finally, the Parallel Factor (PARAFAC) decomposition is utilized to resolve angle estimates. Owing to the reconstructed third-order tensor, angle estimation by the proposed method is more accurate than existing methods and has stronger robustness even though there is a strong correlation between targets. Simulation results verify the advantages of our solutions over its cutting-edge counterparts.