Research On Application Of Stochastic Resonance And Chaos Theory In Weak Signal Detection

With the booming development of the nonlinear science in recent years, especially, the discovery of stochastic resonance and chaotic theory provided two brand new ways to the field of weak signal detection. According to the exited stochastic resonance theory, we cannot obtain the dynamic behavior of a stochastic resonance(SR) system intuitively. In order to reveal the dynamic mechanism of SR, a kind of first-order Duffing equation attractor is analyzed at first, and then the property of nonlinear Duffing equation is studied, based on which the nonautonomous Duffing equation attractor curve is deduced. The output of SR system can be obtained by mapping the input signal on the attractor curve, and the dynamic mechanism of SR is explained by using the mapping method. Analysis of the result indicates that the intrinsic signal can push the system to move along the attractor curve, and the noise can evoke a transition response of the system under the given conditions. Some exited SR weak signal detection methods, such as the parameter-adjustment and damping-adjustment are extended by the proposed dynamic mechanism.Chaotic detection method can be used to pick out the weak sinusoidal signal submerged in noise. Three types of detection properties of this method are proposed in this paper, i.e., sharp change ability of phase diagram, maintenance zone of chaos, and the sufferance with noise of the chaos critical point. After analyzing the factors influencing these three properties, two different chaotic systems are compared according to the three properties. On the other hand, there are few effective ways to construct the chaotic detection system for detecting a continuous frequency band signal. We elaborate two chaotic detection methods, the fixed-threshold-value method and the floating-threshold-value method, and analyze their advantages and shortcomings. On this basis, we propose the multi-frequency-band threshold value transform method to deal with the problems existing in the above two mentioned methods, and design the self-frequency-tracing circuitry accordingly. The operation of this circuitry is validated by hardware experiments. Experimental results show that weak medium-low frequency signals can be detected from the background noise by using this method.