Robust High Data Rate MIMO Underwater Acoustic Communications

摘要

Our research focuses on mobile multi-input multi-output communications over sparse acoustic channels subject to both inter-symbol interference and Doppler scaling effects. Temporal resampling is implemented to effectively convert the Doppler scaling effects to Doppler frequency shifts. We evaluate two practical channel models and compare the performance of the corresponding channel estimation algorithms, namely, the generalization of the sparse learning via iterative minimization (GoSLIM) algorithm and its variation, GoSLIM-V. For symbol detection, we consider the Turbo equalization schemes implemented by the linear minimum mean-squared error based soft-input soft- output equalizer and its low complexity approximation, respectively. The latter provides only slightly degraded detection performance at a significantly lower computational complexity compared to the former. The effectiveness of the considered approaches is verified using the WHOI09, ACOMMIO, and MACE10 experimental results.