Multi-resolution BCS-based approach for DoA estimation

The problem of the estimation of the number and angles of arrival of electromagnetic signals impinging on a linear array of sensors is addressed by means of a multi-resolution approach based on the Bayesian Compressive Sensing (BCS). The BCS is used to retrieve at each step of the multi-zooming strategy the most probable directions of arrival (DoAs) as well as to provide the confidence level of the obtained estimations. The DoAs estimation is then refined by zooming on the spatial regions where the signals have been detected and the zoom velocity is defined according to the confidence on the reliability of the estimations in the previous step. A set of preliminary numerical results is provided in order to assess the performance of the proposed technique.