Direction-of-Arrival Estimation Using a Sparse Representation Based on Fourth-order Cumulant

Sparse iterative covariance-based estimation (SPICE) method is a computational efficient sparse method for direction of arrival (DOA) estimation but has a poor performance in resolution and noise immunity. The high-order cumulant can extend the array aperture and reduce the Gaussian noise. Therefore, this paper proposed an improved SPICE based on fourth-order cumulant, which shares the same features of SPICE but has higher resolution and outperforms in low SNR case. Moreover, its computational cost is comparatively low by distilling the un-redundant data of uniform linear array. Simulations were conducted to validate and evaluate the proposed method.