Adaptive algorithms for blind channel equalization in impulsive noise

• Published in: Signal processing 2022-12, Vol.201, p.108626, Article 108626
• Main Authors: Abrar, Shafayat, Zerguine, Azzedine, Abed-Meraim, Karim
• Format: Article
• Language: English
• Subjects: Adaptive filter; Blind equalization; Engineering Sciences; Generalized Gaussian distribution; Impulsive noise; Maximal-invariance; Signal and Image processing
• ISSN: 0165-1684; 1872-7557
• DOI: 10.1016/j.sigpro.2022.108626

•Exploiting Wijsman’s theorem, we obtain a test statistic serving as an admissible cost function for blind equalization robust to impulsive noise.•We establish the admissibility of the proposed cost function under multipath scenario.•We derive two realizations of adaptive equalizers, and provide simulations on the convergence behaviour of all addressed equalizers.•Using energy conservation, we provide selective-update analysis of one of the proposed equalizers, and obtain a bound for its step-size. The unsupervised adaptive mitigation of intersymbol interference in an additive impulsive noise environment, modeled as generalized Gaussian, is dealt in this work. The theory of statistical invariance, Wijsman’s theorem, is used to develop a maximal-invariant test to discriminate equally-likely pulsed signals against impulsive disturbance leading to an admissible cost function for blind equalization. The cost function is optimized to realize two adaptive equalizers capable of not only mitigating intersymbol interference but also robust to impulsive disturbance. Numerical simulations, obtained on a baseband digital microwave radio system for amplitude-phase shift keying signaling in an additive (generalized Gaussian and symmetric-alpha stable) impulsive environment, confirm the admissibility of the proposed equalizers in terms of robustness and steady convergence.