Robust adaptive DOA estimation method in an impulsive noise environment considering coherently distributed sources

Abstract

Many approaches have been studied for the direction of arrival (DOA) estimation of distributed sources under additive Gaussian noise environments, but these schemes typically perform poorly when the noise is modeled as an alpha-stable distribution. This paper extends the definition of correntropy, which exhibits a robust statistical property under impulsive noise environments, by proposing a novel generalized autocorrentropy (GCO) operator. To improve the performance of the GCO operator, an adaptive kernel size function is deduced for symmetric alpha-stable (SαS) distributed random variables. Since the adaptive kernel size function does not require any empirical parameter of the impinging signals, it is particularly suitable for cases of practical interest. Based on the advantage of the GCO operator, a distributed signal parameter estimator (DSPE)-like algorithm is proposed for the DOA estimation of coherently distributed (CD) sources in the presence of impulsive noise. Comprehensive simulation results show that the proposed algorithm surpasses the existing algorithms concerning estimation accuracy and probability of resolution under the SαS distributed impulsive noise environments. We also conduct comparisons with the existing DOA estimation algorithms under the other two typical impulsive noise conditions.