High-Resolution DOA Estimation in the underwater radiated noise based on Sparse Bayesian Learning

Direction of arrival (DOA) estimation is one of the most important issues in the array signal processing, not only giving the spatial positioning of the targets of interest, and also providing the technical support for the signal enhancement; hence it has a wide range of applications in sonar, radar and communication. This paper focus on high-resolution wideband DOA estimation in the noise radiated from underwater targets of interest in the framework of sparse Bayesian learning (SBL). In the conventional wideband DOA approaches in the sparse reconstruction (SR), the group sparsity between frequencies is used to improve the DOA reconstruction performance. However, this approach ignores correlations between Fourier coefficients in the group and treats them independently. In the paper, we introduce a Toeplitz matrix based on the auto-aggressive model to model these correlations, and propose a novel wideband DOA estimation method by jointly exploiting group sparsity and their correlations of Fourier coefficients within group in the framework of SBL, the proposed DOA algorithm has an enhanced performance over conventional SR based algorithms. The simulation results verify the correctness and effectiveness of the proposed algorithm.