Iterative Channel Estimation and Turbo Equalization for Multiple-Input Multiple-Output Underwater Acoustic Communications

Abstract

This paper investigates a multiple-input multiple-output (MIMO) detector for underwater acoustic (UWA) communications, that uses the improved proportionate normalized least mean squares (IPNLMS) algorithm for iterative channel estimation in turbo equalization. By exploiting the sparse nature of UWA channels, the iterative IPNLMS channel estimator achieves fast convergence and fast tracking by data reuse of training sequence and decision-directed channel reestimation over each turbo iterations. The proposed channel estimation scheme is combined with the low-complexity minimum mean-square-error (LC-MMSE) equalization to detect single carrier MIMO signals in overlapped subblocks without guard intervals. The proposed MIMO detection scheme achieves high performance, high transmission efficiency, and low computational complexity, especially for large MIMO and high-constellation modulation schemes. The experimental results of the undersea 2008 Surface Processes and Acoustic Communications Experiment (SPACE08) demonstrate an order of magnitude improvement of bit error rate (BER) performance over those using block-wise MMSE channel estimation. We have also verified that the channel-estimation based turbo equalizer (CE-TEQ) outperforms the direct-adaptation based turbo equalizer (DA-TEQ) in terms of BER performance and robustness against error propagation.