Study On Weak Signal Detection Based On Bi-Stable Langevin Model

The method of detecting weak signal by bi-stable nonlinear system has been used in many fields for its predominant advantages than traditional methods, such as chemistry analysis, instrumental analysis, and etc. However, using this method in detection of weak signal was limited in adiabatic approximation condition, as much as to say the input signal frequency, signal amplitude and the density of incidental noise must be far less one. This limitation made a restriction of the application of bi-stable system. To broaden the application of bi-stable nonlinear system in weak signal detection, the paper did theoretical and experimental deeply study based on bi-stable Langevin model.The paper analyzed the SNR of the output in stochastic resonance system based on adiabatic approximation condition in bi-stable nonlinear system and compared the SNR with the one obtained by the directly numerical simulation of the Langevin model. The comparison result showed the limitation of adiabatic approximation theory when the frequency of signal is close to zero or too big.A numerical analysis method called four order Runge-Kutta was employed to analyze the action of bi-stable Langevin model and weak signal with big frequency. Our results showed that bi-stable Langevin model can restrain noise and increase SNR when frequency was not very low. The deduction of the moment equation explained the quantitative relationship between output and input of bi-stable system from moment method.The paper introduced two technologies to heighten the frequency of weak signal detected by bi-stable Langevin model. The introduction of sequential step sampling into stochastic resonance heightened the frequency of signal detected by stochastic resonance system to 2 kHz. When the frequency of await signal was relative high, filter technique and bi-stable Langevin model were combined to detect the weak signal with the frequency of up to 500Hz. by bi-stable system directly.The process of detecting weak signal by bi-stable Langevin model was studied and analyzed through numerical simulation and Labview virtual...