Estimation of Doubly Spread Underwater Acoustic Channel via Gram-Schmidt Matching Pursuit

The underwater acoustic channel (UAC) exhibits strongly time delay and Doppler (DD) spread especially when the UAC is rapidly time-varying. These dynamic factors result to a serious impact on communication performance such as Inter-Symbol Interference (ISI). Hence, estimation of complex amplitude, time delay and the Dopplers of the UAC becomes the key part in underwater acoustic communication and is hopeful for improving the performance of equalization. However, the estimation is challenged by multiple factors to be estimated in delay and Doppler dimensions. This study exploits the sparsity of the UAC and develops an estimator via using Gram-Schmidt to find orthogonal bases, which leads to the fast and orthogonal way to select the supports of the dictionaries. The support list of the dictionaries constructed by probe signal can be used for estimating the DD functions from a noisy received signal. Matching Pursuit (MP) and Least Square (LS) methods are used for comparisons. The effectiveness of the proposed method is verified by the experimental data.