Research On Weak Signal Detection Based On Asymmetric Tristable Stochastic Resonance System

In many experimental studies and engineering practices,weak signals are detected when there is a low signal-to-noise ratio.As a nonlinear detection method,stochastic resonance detection method is different from the traditional weak signal detection method.At the same time,it will not weaken the signal characteristics of the useful signal,because it uses the method of transferring noise energy to achieve the purpose of weak signal detection.The current research fields of stochastic resonance systems focus on symmetrical stochastic resonance systems,which are difficult to maintain symmetry in practical applications,and most researches on asymmetric stochastic resonance systems focus on double-potential well systems.As a means to better extract the signal features under strong noise backgrounds,and therefore improve the output signal's signal-to-noise ratio,stochastic resonance systems can be used,an asymmetric tristable stochastic resonance system of combined asymmetry is proposed in this paper.Here is the main work of the paper:(1)The stochastic resonance effect of an asymmetric tristable stochastic resonance system driven by white Gaussian noise is studied.Firstly,through the adiabatic approximation theory,asymmetric tri-stable stochastic resonance system's effective potential function and first pass time are derived,the influence of the parameters on the first pass time is analyzed in order to demonstrate the necessity of optimizing the system parameters adaptively.Secondly,A system is used to detect weak signals in the background of Gaussian white noise,and the system's weak signal processing capability is demonstrated.Effects of stochastic resonance output responsiveness.(2)Combining Gaussian potential well with asymmetric bistable potential well,an asymmetric tristable stochastic resonance system is proposed.In the first step,a system model is established and the Langevin equation is derived;then,under a Gaussian white noise background,the system is simulated to process several types of periodic signals,It is verified that the system has weak signal processing capabilities;finally,system performance index is based on the signal-to-noise ratio output and the average gain of the signal-to-noise ratio,the influence of noise intensity on stochastic resonance's output response is discussed in detail.(3)A particle swarm algorithm optimizes the parameters of the two systems,and compare the input signals before and after optimization to verify the necessity of adaptive optimization of system parameters.Then,under the input conditions of several types of periodic signals in this paper,in the system performance metric,signal-to-noise is used,and the comparison of the adaptive combined asymmetric tri-stable system is compared with the adaptive asymmetric bistable system and the symmetric tri-stable system.The weak signal processing ability of the state system and asymmetric tri-stable system is better.Then,the output signal of the self-adaptive combined asymmetric tri-stable system is compared with the signal reconstructed by the sym wavelet decomposition.Finally,the adaptive combined asymmetric tri-stable system,symmetric tri-stable system,asymmetric bistable system,and asymmetric tri-stable system are used to process the actual bearing fault signal.The measured results prove that the combined asymmetric tri-stable system The system is practical and superior to several other types of adaptive stochastic resonance systems.This paper provides a new option for the practical application of the stochastic resonance method.