Weak Periodical Signal Detection Based on Wavelet Threshold De-Noising and Chaos Theory

DENG Hong-gui,CAO Wen-hui,YANG Bing-chu,MEI Wei-ping,AO Bang-qian
2012-01-01
Journal of Central South University
Abstract:Based on the multi-resolution of wavelet transform and chaotic system having a good immunity to noise and sensitive to weak periodical signal,a new method of weak signal detection was proposed based on the combination of the wavelet threshold de-noising and chaotic system by improving wavelet threshold de-noising and duffing oscillator.This method uses wavelet smoothing effect to the limited discrete processing of the signal that contains noise and uses the scale of the wavelet decomposition to determine the de-nosing depth,and then uses the reconstructed signal as the driving motivation of the perturbation cycle into the chaotic system.The chaotic oscillator array is applied to detect weak signal in noisy background,and Melnikov method is adopted as chaotic criterion.This new method has overcome the blindness to the determination of scale and irrationality choice to the threshold of past method of the wavelet decomposition,as well as the ambiguity to distinguish between the critical state and periodic motion state.At the same time,multi-frequency signal can be detected.The simulation results indicate that this detection method is simple and effective,and the detection precision is high.The minimum detection SNR reaches -100 dB and the frequency error is about 0.04%.The weak signal buried in strong noise detection method is improved.
What problem does this paper attempt to address?