Covariance Based Tensor Reconstruction Approach for the Bistatic MIMO Radar in Unknown Nonuniform Noise

To resolve the problem of unknown nonuniformity of noise that occurs during data signal reception of the bistatic MIMO radar, a covariance-based tensor reconstruction approach is proposed in this paper. After obtaining the tensorial arrangement of the received measurement, a tensor reconstruction model which permits a tensor transformation with an inherent linear transformation matrix is utilized to capture, exploit and get rid of the unknown noise variances. Then the covariance matrix of the obtained received data measurement taken and reduced to a single structure angle estimation matrix problem where a Reduce dimension MUSIC method is adopted to exploit the signal subspace. Contrary to existing methods, the proposed method is simple and robust to unknown nonuniform noise with an improved angle estimation performance. Results from numerical simulation provided demonstrate the validity and effectiveness of the proposed method.