Under-ice acoustic channel estimation method based on adaptive sparsity orthogonal matching pursuit

Based on the stable multipath structure characteristics of the under-ice acoustic channel, a scheme utilizing the preamble signal of the communication frame for channel estimation is adopted. Leveraging the sparse characteristics of the under-ice acoustic channel, an adaptive sparsity orthogonal matching pursuit (ASOMP) algorithm is proposed to address the reliance of the orthogonal matching pursuit (OMP) algorithm on prior knowledge of sparsity. Firstly, the sparsity of the channel is coarsely estimated using the autocorrelation characteristics of the preamble signal, and the simplified dictionary matrix is constructed. Subsequently, based on the coarsely estimated sparsity and the simplified dictionary matrix, the OMP algorithm is employed to obtain a preliminary estimation of the channel impulse response (CIR). Finally, fine estimations of both the sparsity and CIR are performed using residual differences derived from the coarsely estimated CIR. The adopted channel estimation scheme eliminates the need to insert training sequences into the data segment, thereby enhancing the data rate. The proposed channel estimation method enables effective channel estimation without requiring prior knowledge of channel sparsity. Simulation results show that the ASOMP algorithm achieves performance comparable to the OMP algorithm with known sparsity and outperforms the existing sparsity adaptive matching pursuit (SAMP) algorithm under low signal-to-noise ratio conditions. The data processing results from the 14th Chinese National Arctic Research Expedition’s under-ice experiment validate the effectiveness and reliability of the proposed algorithm.