Laser pulse signal waveform recognition based on bidirectional long short-term memory neural network

To address the problems existing in the pulse signal waveform recognition module in the field of laser semi-active guidance, such as difficult signal feature extraction, diverse waveform types, and low computational efficiency, a laser pulse signal waveform recognition method based on a bidirectional long short-term memory (BiLSTM) neural network was adopted. By leveraging the intuitive characteristics of laser pulse signals in the time domain, the recognition and classification of laser signals with different powers and background noise signals were achieved, and the recognition accuracy was higher than 99.7%. The experimental results show that BiLSTM has more excellent network performance compared with the simple long short-term memory neural network. On the premise of ensuring high performance, the anti-interference capability of the laser semi-active guided weapon has been effectively improved.