Time-Domain Adaptive Equalization Technique Based on Bidirectional Superimposed Pilot

To enhance the tracking capability of communication systems for time-varying underwater acoustic channels and improve spectral efficiency, this paper proposes a time-domain adaptive equalization technique based on bidirectional superimposed training sequences. By linearly superimposing training sequences with information sequences, this method eliminates the additional bandwidth consumption associated with conventional interleaved training sequences. To address the characteristics of underwater acoustic channels featuring long delay spread and strong multipath effects, this paper develops an iterative optimization framework for bidirectional adaptive channel equalization and estimation: By superimposing training sequences in both forward and reverse transmission directions, a bidirectional iterative optimization mechanism is constructed to dynamically track channel characteristics. Furthermore, incorporating adaptive channel estimation, a symbol-by-symbol superimposed training sequence interference cancellation scheme is proposed to eliminate training-induced interference in the time domain. Finally, the interference-canceled results from both directions are jointly processed. After decoding and reconstruction, the output is fed back to the channel estimation module to enhance estimation accuracy, while the soft symbol information from Turbo equalization is iteratively fed back to the equalizer for optimization. Simulation results demonstrate that, compared to conventional interleaved training sequence methods, the proposed approach significantly improves the performance of underwater acoustic communication systems.