Cross-Correlation-Based Complex-Valued Sparse Deconvolution Beamforming for Active Sonar System in Multipath Environment

Direction-of-arrival (DOA) estimation is fundamental for marine applications, but it faces significant challenges in multipath environments where the received signals are often coherent. While several algorithms exist for narrowband signals, they struggle with wideband signals widely used in the active sonar system. To address these issues, a cross-correlationbased complex-valued sparse deconvolution beamforming (CCSDB) is proposed in this paper. To fully utilize the known information of source signal, conventional DOA estimation model is developed with cross-correlation function, thus the output SNR is maximized, and the robustness of wideband beamforming is enhanced. To obtain high-resolution DOA estimation at coherent case, a complex-valued deconvolution algorithm with sparseness and continuity constraints is applied as a post processor in the developed model, thereby suppressing the main-lobe width and sidelobe level. Numerical simulations and underwater experiments validate the algorithm's effectiveness.