Combined boosted variable step-size affine projection sign algorithm for environments with impulsive noise

The affine projection sign algorithm (APSA) is widely used to alleviate the impact of impulsive noise on weight updates for adaptive filters. When using the conventional APSA, the sign-based approach tends to degrade the rate of convergence; however, this can be remedied by adopting a variable step-size (VSS). Nonetheless, the tracking performance of the APSA tends to be poor in situations that involve identifying rapid changes in the system model. Inspired by the boosted adaptive filtering framework, we propose using multiple component filters (weak learners) to calculate the local boosted VSS. Specifically, this involves combining the boosted VSSs of the first and last stages of the component filters. In simulations, the proposed combined boosted VSS (CBVSS) method was shown to outperform comparable works in terms of convergence rate, even in situations where the unknown system model changes abruptly.& COPY; 2023 Elsevier Inc. All rights reserved.