| Stochastic resonance is a new technique for weak signal detection.In this thesis,the background and significance of stochastic resonance are discussed.The stochastic resonance method is used to detect weak signal.The mathematical model of bistable system is studied.The classical stochastic resonance method is supplemented and improved to enhance the detection performance of weak signal.The signal to noise ratio(SNR)of adaptive system is used as fitness function.The optimal weak signal detection results are obtained by adjusting the parameters of the noise and bistable system.The method of stochastic resonance is simple and fast.The fast detection of signals can be achieved in a short data set.In the special structure of the bistable model,the stochastic resonance can transform the partial energy of the noise into the excitation of the signal,which makes the signal feature more prominent.Compared with other detection methods,suppressing noise is not the main means in this method.This method overcomes the disadvantage of the loss of useful signal when the noise is suppressed,which provides a new way for signal detection.Stochastic resonance can shorten the time of weak signal detection.This method can obtain better output results by using fewer data sets.Classical stochastic resonance theory can not detect unknown frequency signal.The amplitude of the signal can not be accurately measured by this method.At the same time,SNR gain is not enough.In order to solve these problems,an improved method is put forward.The main contents of this thesis are as follows.(1)The background and significance of the signal detection of stochastic resonance method are discussed.The stochastic resonance theory is discussed in the domestic and foreign research results.The stochastic resonance phenomena in various experiments are listed.The applications of stochastic resonance in mechanical fault,medical and biological sciences are discussed.(2)The theoretical basis of stochastic resonance is analyzed by Langevin equation,adiabatic approximation theory and linear response theory.The measurement indexes of signal detection are analyzed.The establishment of the bistable model of stochastic resonance is discussed.The nonlinear characteristics of Langevin equation are analyzed qualitatively.The detection method of stochastic resonance signal based on bistable system is discussed.The influence of bistable detection model on stochastic resonance is analyzed.(3)In this thesis,the method of detecting the frequency of weak periodic signal with stochastic resonance is studied.Particle swarm optimization algorithm is used to adjust the external noise intensity and the parameters of stochastic resonance system.The effects of the detection results by adjusting the noise intensity and the system parameters are verified.At last,the model of cascaded stochastic resonance and the adaptive stochastic resonance method are proposed.(4)If the input SNR is in a certain range,the stochastic resonance can be used to deduce the amplitude of the detected signal according to the output SNR and the gain of the SNR.However,this method can not detect the signal amplitude accurately.In order to solve this problem,a new method is proposed,which is based on the combination of stochastic resonance and chaotic oscillator.Firstly,the principle of weak signal detection based on chaotic oscillator is described.The abrupt change of the chaotic oscillator at the critical point is analyzed.The effects of noise and other frequency signals on the Duffing oscillator are discussed.Finally,the method of detecting signal based on stochastic resonance and chaotic oscillator is verified by simulation.(5)The theory and model of aperiodic signal detection based on stochastic resonance are studied.The simulation results show that the pulse signal can be detected by stochastic resonance.The influence of the peak height of the pulse signal on the detection results is analyzed. |
