DOA Estimation Based on Approximate l0-Norm Sparse Reconstruction Under Alpha Stable Distribution noise

The article delves into the critical area of direction of arrival (DOA) estimation in array signal processing, highlighting the limitations of traditional subspace-based algorithms in resource-constrained environments. It introduces compressed sensing (CS) technology as a viable solution, leveraging the sparsity of target signals in the spatial domain. The paper categorizes existing CS-based DOA estimation algorithms into greedy algorithms, sparse Bayesian learning (SBL) algorithms, and convex optimization algorithms, each with its own strengths and weaknesses. A significant focus is placed on the grid mismatch problem, which occurs when the target signal deviates from the discretized spatial domain, leading to performance degradation. The article proposes an off-grid approximate l0-norm algorithm (BP-EOGSL0) that utilizes a bounded nonlinear function (BNF) and phased fractional lower-order moment (PFLOM) to suppress Alpha noise, a type of non-Gaussian impulsive noise. This algorithm constructs an exponential family distribution (EFD) function to approximate the l0-norm, enhancing the accuracy of sparse reconstruction. The paper also addresses the grid mismatch effects through a first-order Taylor expansion of the steering vector, introducing off-grid deviation into the sparse DOA model. Simulation results demonstrate the high estimation accuracy and success rates of the proposed algorithm under impulsive noise conditions. The article concludes with a discussion on the computational complexity of the algorithm and potential future improvements, making it a comprehensive resource for those interested in advancing DOA estimation techniques in challenging noise environments.