Blind channel estimation and phase ambiguity elimination in MIMO-OFDM UWA communications

Underwater acoustic (UWA) channel is bandwidth-limited owing to the low sound speed, severe time- and frequency-selective fading. We believe that improving the spectral-efficiency is a decent approach to increase the capacity of UWA systems. In this paper, we introduce a highly spectrum-efficient MIMO-OFDM system under UWA channel that uses blind channel estimation and elimination of phase ambiguity induced in equalization. We employ classic subspace algorithm (SPA) as the blind channel estimation method without the training sequences and pilot tones. To maximize the performance of SPA, a resampling-based solution to mitigate the channel Doppler shift is applied at the receiver. Considering the phase ambiguity introduced inherently in SPA, we use an interleaved symbol mapping schemes to eliminate the ambiguity, which achieves no rotational symmetry in the combined signal constellation. Numerical simulations verified that the phase error could be corrected without any data aiding and the spectrum efficiency of MIMO-OFDM UWA communications is greatly improved compared with pilot-based method.