Snapshot-deficient Off-grid Direction of Arrival Estimation for Underwater Scenarios via Dictionary Refining Elastic-Net Framework

This paper presents a new method for sparse direction of arrival (DOA) estimation, aimed at resolving off-grid challenges in scenarios with limited snapshots, particularly in single-snapshot conditions. Traditional single-snapshot DOA sensing approaches, relying on a predefined grid and potentially incorporating the Least Absolute Shrinkage and Selection Operator (LASSO) algorithm for sparse recovery, often exhibit suboptimal performance. Estimation of off-grid directions serves to alleviate the difference between observed target bearings and assumed grid points. Increasing the density of grid points can improve estimation accuracy, but it also significantly increases computational complexity. To address these challenges, this study suggests a sparse recovery approach based on the Elastic-Net model. The wideband integrated dictionary is incorporated within the iterative sensing matrix to address computational complexities and resolve grid mismatch issues. The simulation and experimental results confirm that the proposed algorithm outperforms traditional sparse signal recovery techniques in terms of root mean square error and the probability of exact recovery, effectively reducing the negative impact of grid mismatch.