A simple sparse channel estimator for underwater acoustic channels

Underwater acoustic channels exhibit very long delay spreads, but with limited multipath. To this end, channel estimation for sparse underwater acoustic channels is examined. Based on the on-off keying techniques developed in [1], a new iterative scheme for sparse channel estimation is provided. The method is of lower complexity than that of [1], exploiting features of underwater acoustic channels. The strategy iterates between zero-tap detection and structured channel estimation. The new sparse channel estimation scheme is compared to the Matching Pursuit method. It is shown that the proposed algorithm outperforms the latter over realistic underwater acoustic channels and has comparable computational complexity. Furthermore, the new estimation method does not require channel order information, in contrast to Matching Pursuit.