Estimation of Underwater Acoustic Channel via Block-Sparse Recursive Least-Squares Algorithm

Underwater acoustic communication suffers from the serious multipath effect and fast time variation due to the sophisticated ocean environment. General channel estimation schemes have limited performance in underwater acoustic communication because of the lacking exploitation of the underwater acoustic channel (UAC) inherent property. In this work, we address a new approximate mixed ℓ
_2,0
-norm, and based on that norm, we develop a block-sparse recursive least-squares (BS-RLS) algorithm for UAC estimation which takes advantage of the underlying block-sparse structure of UAC. The simulation result shows that the proposed BS-RLS algorithm improves the channel estimation quality under block-sparse condition.