Detection of Weak signal and Fusion of Neural Network Data in Ultrasonic System Using Chaos Theory

摘要

Within the frame work of ultrasonic detecting system, a theoretical method was proposed for extracting weak signals from the background of strong noise using the chaos character first. Taking the advantage of sensibility of the chaos vibrator to valuable signals and the strong immunity to noise, the detected signals can be made available from all kinds of complex noises. Multi-sensor data fusion technology is used for information processing, using artificial neural network. The best structure of network was established for the accuracy and training rate required for neural network.