Noise Suppression and Signal Recovery Algorithms in Underwater Acoustic Communication Systems

This paper studies the noise suppression and signal recovery algorithms in underwater acoustic communication systems. The underwater acoustic channel characteristics and noise statistics are analysed through the measured data in the South China Sea, and the noise suppression algorithm based on the improved U-Net (with residual connectivity and attention mechanism) and the signal recovery algorithm based on the bidirectional long and short-term memory network (BiLSTM) are designed. Compared with the traditional methods such as Kalman filtering and wavelet denoising, the algorithm in this paper can additionally improve the signal quality by 3-4dB under the condition of -10dB signal-to-noise ratio (SNR). The DSP+FPGA experimental platform is built for sea trial demonstration, and the results show that the scheme can improve the signal quality by 8.5dB under the condition of -10dB SNR, and achieve the reliable communication with the BER better than 10<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-4</sup> at the communication distance of 2km. The system maintains more than 92% of communication success rate under 6 levels of sea state, and the total processing delay is less than 1.8ms, which verifies the practicality of the algorithm.