Robust diffusion recursive adaptive filtering algorithm based on lp-norm

Abstract

A robust diffusion adaptive filtering algorithm, called the diffusion recursive least l p -norm (DRLP), is developed for distributed estimation over network. The new algorithm aims at recursively minimizing the l p -norm of error, and can offer a more stable and robust solution than traditional adaptive filtering schemes based on minimization of the squared error, such as the diffusion recursive least squares (DRLS) algorithm. Simulation results show that the proposed DRLP can outperform several state-of-the-art methods especially when the network is disturbed by impulsive noises.