Sparse channel estimation for underwater acoustic communication channels

Abstract

In underwater communication channels, acoustic arrivals often form a cluster structure. The number of significant taps is limited although the length of the channel can be very large. As shown in the previously reported research, channel estimation can benefit the performance of the coherent communication and promise to support high data rate and reliable underwater acoustic communication. Since underwater communication channels are sparse in time domain, channel-estimation algorithms can have lower complexity and higher accuracy by exploiting the sparse property. Various sparse channel estimation techniques, including threshold recursive least-squares (RLS), matching-pursuit, and on-off keying detection algorithms, are investigated in underwater communication channels. Their effectiveness and complexity are compared using communication data measured from the Kauai experiment, July 2003. Their impacts on the communication performance are also presented.