ICSV26

In many instances, the ambient ocean noise is impulsive in nature. Classical channel estima-tion methods optimized for Gaussian noise suffer serious performance degradation even when the noise is slightly impulsive. This paper deals with the problem of sparse underwater acoustic (UWA) channel estimation under impulsive noise environments with non-Gaussian distribution. Channel estimation is realized based on the recursive least p-norm (RLP) algorithm which has been proved to be robust under impulsive noise environment. L10 or L20 norm is incorporated to the cost function to take advantage of the group sparse nature of UWA channel. Simulation results show that the proposed RLP can resist the influence of the impul-sive noise and L10-RLP and L20-RLP has better estimation performance regardless of Gaussian or non-Gaussian noise when group sparse channel encountered.