DOA estimation for wideband signals via the Homotopy approach

Abstract

Compared with the conventional subspace-based methods, direction-of-arrival (DOA) estimation from the sparse recovery perspective renders higher resolution capability and lower data requirement. When wideband signals are handled, however, straightforward adoption of such an idea is no longer reasonable, due to the diversity of estimation at each frequency bin. In this paper, we address the problem of DOA estimation for wideband signals according to the weighted subspace fitting (WSF) criterion, which is solved by the Homotopy method. The proposed algorithm holistically exploits the joint sparsity feature embedded in and among multiple frequency bins, and does not require any difficult selection of hyperparameters. Simulation results illustrate the superior performance achieved in the proposed algorithm.