Abstract This paper addresses the degradation in direction of departure (DOD) and direction of arrival (DOA) estimation performance caused by array mutual coupling in bistatic MIMO sonar systems. A joint DOD-DOA estimation algorithm based on smoothed Atomic Norm Minimization (ANM) is proposed. The method mitigates the effects of array mutual coupling by applying a decoupling smoothing matrix to preprocess the received signals. The processed signals exhibit a structure resembling line spectral signals, which allows for the formulation of an ANM optimization problem. By exploiting the inherent sparsity of the signals, the proposed algorithm effectively reconstructs a high signal-to-noise ratio (SNR) covariance matrix, even with a limited number of snapshots, enabling accurate estimation of DOD and DOA. Simulation results validate the algorithm’s performance under mutual coupling conditions in MIMO sonar transmit and receive arrays. The algorithm maintains high estimation accuracy even with a limited number of received snapshots, making it suitable for rapid target localization in real underwater detection applications. Compared to existing gridless sparse recovery methods for covariance matrix estimation, the proposed algorithm achieves lower computational dimensions and faster execution. Additional simulation experiments further demonstrate the effectiveness and superiority of the proposed method under varying SNR and target angle separation conditions.
Gao et al. (Mon,) studied this question.