Sparsity Model and Performance Analysis of DOA Estimation with Compressive Sensing

The method of Direction-Of-Arrival (DOA) estimation of array signals based on Compressive Sensing (CS) theory has advantages such as fewer snapshots requirement and the capacity of dealing with the coherent sources. Exploiting the CS theory on DOA estimation, one of the key issues is to construct the sparsity model of source signals. This paper proposes the systemic analysis about how the way of space-partition affects the performance of DOA estimation, and presents a new optimal sparse reconstruction model based on space-partition with equal sine interval through the analysis about coherence. The theoretical result shows that the reconstruction model based on the manifold matrix with equal sine interval is the best model in the practical application. Finally the experiments compare the reconstruction performance of the manifold matrix with equal sine interval with that of the manifold matrix with equal angle interval. This paper provides the experiment results about the performance of signal reconstruction and DOA estimation, respectively. The advantage of the presented sparsity model is verified.