Research On Weak Signal Detection Based On Stochastic Resonance And Wavelet Analysis

Weak Signal Detection(WSD)technique is widely used in various fields,such as wireless telecommunications,fault detection,radar system,geological signal processing and biomedical signal processing.The object of WSD is a weak amount that can not be detected by conventional and traditional methods.With the development of science and technology,the demand for WSD is increasingly urgent.That is to say,WSD is an important approach for developing high-tech and exploring nature.Stochastic Resonance(SR),synergetic effects among nonlinear systems,input signals and noise,enhances weak signals by transferring some energy of noise to weak signals and overcomes the harmful effect of noise in traditional WSD.Therefore,SR is a hotspot in WSD technology.Based on a comprehensive introduction of basic theories and achievements of SR,this thesis studies the application of them in the detection of SR weak signals from three different perspectives: the noise,control methods and input signals.The main works and contributions of this thesis are summarized as follows:(1)The multi-stable SR in WSD method under color noise is studied.Firstly,the generation method of color noise and the system model of SR in multi-stable system are analyzed.Then,the multi-stable SR is deeply studied by using the Signal-to-Noise Ratio Improvement(SNRI)as a measurement index.Finally,experimental results demonstrate that the system model has a significant detection effect on weak signals and can effectively extract the characteristic frequency information of weak signals.(2)In order to improve the ability of SR to process weak signals,the multiscale noise tuning SR is researched.Firstly,a multiscale noise tuning SR based on Wavelet Packet Transform(WPT)is proposed in this thesis.According to the multi-resolution of wavelet transform,the input signals are adjusted to be a multi-scale component by WPT.Then,the wavelet coefficients are processed and the signals are synthesized.Finally,the simulation shows that the proposed SR method overcomes the limitations of small parameter requirement of the classical SR,and achieves the detection of target signals.(3)The impact signals detection method for second-order enhanced SR is studied.Firstly,the system model and output(Signal-to-Noise Ratio)SNR of second-order enhanced SR are analyzed in this thesis.Then,combining the kurtosis and similarity index,a new characteristic coefficient of impact signals is constructed,which not only takes account of both advantages,but also avoids the missing detection for kurtosis as a measurement index of SR.Finally,the simulation results show that the proposed method is effective and feasible,overcomes the weak signals suppressed by the traditional SR processing method,and realizes the effective extraction of impact components.