Auto-weighted least squares based algorithm for direction finding in heavy-tailed noise
Jin He, Zhong Liu · 2005
We present a novel direction finding algorithm based on least squares error in the presence of heavy-tailed noise. We develop the weighted least squares error using array data only, which we call auto-weighted least squares. We propose a unity snapshot infinite norm weighted scheme. The auto-weighted least squares is itself standard least squares. In simulations, we model the noise as an /spl alpha/-stable distribution, and the performance of the proposed algorithm is superior to related algorithms.