부산대학교 도서관

Adaptive turbo equalization (ATEQ) based on the normalized minimum symbol error rate (NMSER) criterion demonstrates powerful detection performance in underwater acoustic (UWA) communications. However, the long delay spreads of the UWA channels usually lead to very high-order channel equalizers, and result in very slow convergence with massive turbo iterations with ATEQ. In this paper, we design a matching-pursuit based ATEQ (MP-ATEQ) algorithm. MP-ATEQ is based on NMSER criterion and updates the taps of the equalizers in the matching-pursuit manner. It leverages the sparse nature of the UWA channels to improve channel equalization, and can lead to a lower BER without initial settings of the taps. And the MP-TEQ can work with block-wise updates, therefore can easily deal with the high-order equalizers with only a few turbo iterations. Simulation and experimental results show that the proposed MP-TEQ outperforms other NMSER-based algorithms in UWA channels without additional complexity cost.