A Robust Direct Position Determination Method Based on Block Sparse Bayesian Learning in the Presence of Unknown Mutual Coupling

In this paper, a block sparse Bayesian learning (BSBL) based robust direct position determination (DPD) method is proposed for locating multiple emitters in the presence of unknown mutual coupling in distributed sensor arrays. This method initiates by conducting a matrix transformation on the actual steering vector, effectively segregating the position parameter from the mutual coupling coefficient. Besides, the BSBL algorithm is designed to iteratively update the hyperparameters. Ultimately, the position of the radio emitter is estimated based on the mean value of the posterior distribution derived from the reconstruction process. At the same time, a mutual coupling coefficient estimator is provided based on the principle of subspace, which can achieve more accurate localization by compensating the mutual coupling coefficient in the sparse dictionary. Simulation results verify that the developed BSBL scheme outperforms the state-of-the-art solution in terms of localization accuracy and robustness to unknown mutual coupling.