Successive weight update sparse Bayesian learning based receiver for multiuser underwater acoustic communications

This paper proposes a single-carrier multiuser (MU) receiver for MU underwater acoustic communications with time-varying and strong multiple-access interference. The receiver integrates soft successive interference cancellation, successive weight update sparse Bayesian learning channel estimation based on approximate message passing (SWUSBL-AMP), and an improved channel update decision. After the first iteration, the improved channel update decision first estimates the channel of each subblock using the interference-cancelled signals from the previous iteration, avoiding the mismatch of time-varying channels. By running the factor graph serially, the SWUSBL-AMP channel estimation algorithm enables the factor graph of the current subblock to utilize the messages updated by the previous subblock, and update the correlation weighting factors, to accelerate the convergence speed and further improve the channel estimation accuracy. Using the experimental data collected in the South China Sea in 2023 for processing, the results demonstrate that the bit error rate performance of the proposed receiver is better than that of other receivers of the same type. Moreover, we also verify the complexity of the proposed SWUSBL-AMP algorithm can be nearly an order of magnitude lower than that of other algorithms.