Detection Of Weak Signal Based On Chaos And Neural Network

The detection weak signal is an important and pivotal technology in modern detection science. The identity of chaos and neural network can provide a new way and method to the detection weak signal.The weak signal measuring method based on chaos oscillator which is introduced in this paper is a new research direction of rising in recent years. This method is based on the study in chaos dynamics. The weak signal can be detected by chaotic Duffing oscillator owing to its property of sensitive dependence on parameter. This method owes the immunity from noise and can magnify the useful signal. And the detection weak signal from the strong noisy background becomes possible. In the emulation experiment, we use the white noise and color noise to be the background to validate the correctness of this method. We can also preview the prospect which is used in the detection of weak signal based on low system SNR.The theory of phase space reconstruction is introduced and the method of chaotic time series prediction is argued in detail. For the intellective character of neural network, we can combine the chaos signal and neural network; especially use the RBF neural network's nonlinear approach character, and then we can actualize the detection of weak signal from the strong noisy background. This paper expatiates the network character and studies arithmetic of RBF neural network. Also the emulation experiment validates the feasibility of this method, analyses the affect from parameters and reason of error.