Sparse Adaptive Matching Tracking Method for Underwater Acoustic Channel Estimation Based on Message Passing

‍ ‍Multipath effects in underwater acoustic communication can distort the received signal， thus resulting in severe intersymbol interference in the time domain and frequency-selective fading in the frequency domain. To address the problem of sparse multipath channel estimation in underwater acoustic communication， this paper proposes a sparse adaptive matching pursuit （SAMP） algorithm based on generalized approximate message passing （GAMP）. The SAMP algorithm is a type of greedy algorithm that combines bottom-up and top-down approaches. It approximates the sparsity of the observed signal by iteratively increasing the size of the support set via an atom-matching test， thereby eliminating the requirement for prior information pertaining to sparsity. The GAMP algorithm， which is an extension of the early approximate message passing method， features low complexity and good adaptability. By replacing the least-squares operation with the GAMP sub-iteration algorithm within the support set obtained by the SAMP algorithm， high-precision estimation of the channel amplitude information can be achieved. Since the GAMP algorithm involves only scalar operations and performs only precise channel calculation in the final iteration of SAMP， the computational complexity of the algorithm is reduced significantly. By simulating the algorithm based on the 2024 Western South China Sea acoustic vortex survey experimental data and applying it to underwater acoustic OFDM communication systems， we confirm that the GAMP-SAMP algorithm outperforms the SAMP algorithm in terms of channel estimation and offers significant advantages in practical applications.