A novel correntropy based DOA estimation algorithm in impulsive noise environments

• Published in: Signal processing Vol. 104; pp. 346 - 357
• Main Authors: Zhang, Jinfeng, Qiu, Tianshuang, Song, Aimin, Tang, Hong
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.11.2014; Elsevier
• Subjects: Algorithms; Applied sciences; Arrays; Computer simulation; Correntropy; Detection, estimation, filtering, equalization, prediction; Direction of arrival; Exact sciences and technology; Fractional lower order statistics; Impulsive noise; Information, signal and communications theory; Miscellaneous; Monte Carlo methods; Noise; Signal and communications theory; Signal processing; Signal, noise; Statistics; Symmetric alpha-stable process; Telecommunications and information theory; Correntropy; Direction of arrival; Fractional lower order statistics; Impulsive noise; Symmetric alpha-stable process; Performance evaluation; Monte Carlo method; Parameter estimation; Direction-of-arrival estimation; Array signal processing; Similarity; Alpha stable process; Measurement sensor; Signal estimation; Background noise; Algorithm; Pulse noise; Lower order statistics; Signal processing; Numerical simulation; Random variable
• ISSN: 0165-1684; 1872-7557
• DOI: 10.1016/j.sigpro.2014.04.033

The Direction of Arrival (DOA) estimation under impulsive noise environments remains open in the field of array signal processing. Inspired by the advantage of correntropy which exhibits a ‘robust statistics’ property, this paper proposes a new operator, namely, the correntropy based correlation (CRCO), for independent and identically distributed (i.i.d) symmetric alpha-stable (SαS) random variables. We define the CRCO based matrix for the array sensor outputs and show that it can be applied with MUSIC to estimating DOAs in the presence of complex symmetric alpha-stable noise. The comprehensive Monte-Carlo simulation results demonstrate that CRCO–MUSIC outperforms the existing fractional lower order statistics (FLOS) based MUSIC algorithms especially in highly impulsive noise environments. •We model the impulsive noises in array signal processing as complex SαS processes.•We examine the inherent robustness to outliers of correntropy.•We propose the statistics CRCO (correntropy based correlation) for SαS random variables.•We define the CRCO based matrix for the array sensor outputs.•We apply the CRCO based matrix with MUSIC to obtain DOA estimates.