Estimation and Equalization of Rapidly Varying Sparse Acoustic Communication Channels

A method for estimating rapidly time-varying and sparse channels is presented. The impact of estimation performance on equalization of phase coherent communication signals is investigated. To address both the rapid channel fluctuations and the channel sparse structure, the delay-Doppler-spread function representation of the channel is adopted. The delay-Doppler-spread function may be considered as a first-order approximation to the rapidly varying channel in which each channel component is associated with Dopplers that are assumed constant or slowly varying over a certain averaging time. The sparse structure of the delay-Doppler-spread function is then exploited by sequentially choosing the dominant components that minimizes a least squares error. Comparison is made between nonsparse RLS channel estimation, sparse channel impulse response estimation and estimation using the proposed approach. The performance of channel estimation as well as the subsequent equalization based on those channel estimates are demonstrated for highly dynamic channels using experimental data