Joint Channel and Impulsive Noise Estimation in Underwater Acoustic MIMO-OFDM Systems Based on Parallel SBL Algorithm

Impulsive noise (IN) can significantly degrade the performance of underwater acoustic communication systems. In this paper, we propose a parallel sparse Bayesian learning (P-SBL) algorithm for joint channel and impulsive noise estimation in underwater acoustic (UWA) multiple-input-multiple-output orthogonal frequency-division multiplexing (MIMO-OFDM) systems. The proposed P-SBL algorithm, whose measurement matrix is constructed using a non-orthogonal pilot pattern, can simultaneously estimate the impulsive noise and channel coefficients from multiple transmitters. Simulation results show that the proposed P-SBL algorithm achieves higher channel estimation accuracy and lower system bit-error-rate (BER) compared with existing approaches.