A CS-based strategy for the design of shaped-beam sparse arrays

The problem of synthesizing maximally-sparse linear arrays with complex excitations is solved through a numerically-efficient approach based on the Bayesian Compressive Sampling (BCS). The array design problem is re-cast in a probabilistic framework, and a fast relevance vector machine (RVM) is employed for the computation of the optimal layout and associated complex weights. A preliminary numerical validation is presented to assess the potentialities and limitations of the proposed approach.