Underwater Acoustic Channel Estimation via Basic-CS and Modified-CS using 2-D Frequency Sampling

This paper proposes a new channel estimation scheme for underwater acoustic (UWA) communication in a two-dimensional (2D) frequency domain. The non-uniform sparse channel in the delay-Doppler domain is exploited to perform the sampling across different frequency ranges. Compressed sensing (CS) is used to recover the sparse signal. Basic-CS, which is well known as l
₁
-norm, channel estimation is investigated. Specifically, we reformulate the slowly varying channel tracking of 2D in the Fourier domain. Modified-CS is used to solve the convex relaxation problem of the signal that is sparsest outside of the "known" part. CS and modified-CS are also observed in terms of variable density sampling in the different Doppler frequency range. Simulation results are presented on the prior estimated channel of the SPACE08 experiment for the validation of the presented scheme.