Robust Direction-of-Arrival Estimation Algorithm Based on W Test Statistics

Considering the problem that the performance of algorithms based on conventional second order cyclic statistics degrades in the impulsive non-Gaussian noise which exists widely in practical environments, the signal samples contaminated by impulsive noise are taken as outliers and W statistics for Gauss test based on robust statistics theory is adopted to adaptively trim the impulsive samples, and then a new estimated cyclic correlation expression is formulated. Consequently, a novel robust DOA (direction of arrival) estimation is obtained. The proposed algorithm needs no any prior knowledge of the noise probability density function and can be applied to all types of non-Gaussian noise model situations. Computer simulations are presented to illustrate the superior efficiency of the novel approach to the existed robust algorithms under different non-Gaussian noise environments.