Error performance of equalizers based on sparse and dynamic channel estimation

Abstract

For broadband communications in the very shallow water environment, it has been shown that both model-based dynamic channel estimation and sparse channel estimation are capable of reducing the signal prediction error by several dB [W. Li and J. C. Preisig, J. Acoust. Soc. Am. 118, 2014 (2005)]. Dynamic channel estimation is based on explicit modeling of the rapid channel variations; sparse channel estimation exploits the channel structure formed by the resolved multipath arrivals. In this paper, the error performance of the equalizer based on these channel estimators are analyzed, demonstrated through experimental results and compared with the performance of equalizers that do not explicitly model and exploit the channel dynamics and structure. [Work supported by ONR Ocean Acoustics.]