Robust Relaxation for Coherent DOA Estimation in Impulsive Noise

摘要

In this letter, we consider the coherent direction-of-arrival estimation problem in impulsive noise. An l(p)-norm-based variant of the classical relaxation technique is proposed to tackle this problem. The proposed method successively minimizes the cost function along block coordinate directions. Specifically, at each iteration, only one block of the signal component is updated, while the remaining blocks are kept fixed. Then, instead of solving each block exactly, the proposed method optimizes the parameters in the block iteratively by solving a surrogate function that upper bounds the l(p)-norm. Numerical results show that the proposed scheme offers substantial performance improvement over the state-of-the-art algorithms.