A Robust Coprime Array Interpolation Method for DOA Estimation in Unknown Nonuniform Noise
Jiawen Yuan, Gong Zhang, Xinhai Wang, Fangqing Wen, Yu Zhang, Changyun Qi · 2021 CIE International Conference on Radar (Radar) · 2021
The most existing coprime array interpolation-based algorithms have the potential to increase the degree of freedom for direction-of-arrival (DOA) estimation. However, these algorithms are modeled on the Gaussian white noise and do not consider the possibility of nonuniform noise. To eliminate the unknown nonuniform noise, we propose a robust coprime array interpolation method in this paper. First, the vectorization is performed on the covariance matrix to obtain the difference co-array with overlapping sensors. Next, we develop a robust interpolation technique to optimize the signal model of the nonuniform virtual array and fill in the holes to convert it into a contiguous virtual uniform linear array (ULA). We then divide the virtual ULA into overlapping subarrays and reconstruct the Toeplitz covariance matrix through the denoising constraint. This constraint can alleviate the impact of nonuniform noise to ensure the robustness of the reconstruction. Finally, the DOA estimation is resolved through the MUSIC algorithm. Numerical experiments validate the superiority of the proposed algorithm.