Deep Learning Based Interference Identification and Suppression Method for Maritime Shortwave Communications

Maritime shortwave communications face interference challenges that affect reliability. This study proposes a deep learning-based method for interference identification and suppression using a two-channel convolutional-long and short-term memory network combined with adaptive filtering and waveform reconstruction. Experimental results show 94.3% recognition accuracy at -5dB to 15dB SNR, 76.2% reduction in average BER, and 128.7ms processing delay, significantly improving anti-jamming capability and communication reliability in maritime shortwave systems.