An improved least square channel estimation technique for OFDM systems in sparse underwater acoustic channel

Underwater acoustic (UWA) communication channels are known to be sparse in nature because of very large delay spread and limited number of significant multipath components. In this paper, we propose an improved least square (LS) based technique for sparse channel estimation applicable to orthogonal frequency division multiplexing (OFDM) systems for UWA communications. Through computer simulations, we demonstrate that the proposed improved LS based sparse channel estimation algorithm outperforms the conventional LS estimation technique by deciding significant channel taps adaptively, based on a predetermined threshold. The mean square error (MSE) and the bit error rate (BER) performance of the proposed improved LS estimator has been observed to be significantly better than that of the conventional LS estimator. Specifically, the results reveal an improvement of approximately 4 dB in SNR at a BER of 10-2for the proposed improved LS estimation technique as compared to the conventional LS estimator.