Adaptive estimation of sparse channel based on modified RLS for coherent underwater acoustics communications

Abstract

A new channel estimation approach used in coherent underwater acoustic communication system is presented in this paper. The proposed method consists of two steps: (1) obtain the initial channel estimation based on least square (LS) and the sparse characteristics of underwater acoustic channels; (2) track the channels using the modified recursive least squares (RLS). In the first step, the threshold value to set all coefficients below it to zero is deduced. In the second step, a modified RLS based on the idea of RLS and Kalman filter terminology is proposed. Experimental results show that the proposed method can track underwater acoustic channels well. The lower symbol error rate (SER) and higher output SNR can be achieved.