Soft Direct-Adaptive Turbo Equalization for MIMO Underwater Acoustic Communications

A Soft Direct-Adaptive Turbo Equalization (Soft DA-TEQ) is proposed for robust MIMO UWA communications. The proposed Soft DA-TEQ is derived under on the general frame of Expectation Maximization (E.M.) algorithm, which utilize the a priori soft decisions from the decoder to direct the equalizer's coefficients adaptation. The a priori soft decisions directed adaptation effectively lower the error propagation, which often cause catastrophic performance degradation in Hard DA-TEQ. Besides, in each iteration of the E.M. based soft adaption, the reliability of the symbol estimation is updated and used in the next iteration of soft adaption, which further enhances the convergence of the DA-TEQ. Experimental results show that the proposed Soft DA-TEQ is capable of robust detection in fast time varying MIMO UWA channels, even with low pilot overhead.