Noise Reduction of Torpedo Ultrasonic Fuze Signal Processing Based on Self-Adaptive Wavelet Neural Network

To remove the noise signal effectively in processing torpedo ultrasonic fuze signal, wavelet theory and artificial neural network are introduced simultaneously, and the construction method of self-adaptive wavelet neural network is given. Better fitting signal is achieved by adopting orthogonal Daubechies wavelet group and optimizing the scale parameter. Simulation result shows that this wavelet neural network can effectively suppress noise from ultrasonic fuze signal with less distortion, leading to reduction of target misjudgment and misdetection of torpedo.