The Research Of Stochastic Resonance In Two Kinds Of Threshold Network Systems

In signal processing, when transmitting the signal in the system, the noise present in the systemwill interference in the transmission of signal, and the performance of the system is deteriorated bynoise. However, the presence of noise will become a favorable factor for signal transmission bystochastic resonance. Noise along with signal entering a nonlinear system, with a non-zero noiseintensity, the system’s response to the input signal is optimal, thereby the performance of system isimproved.This paper discusses the phenomena of stochastic resonance in sum threshold networks basedon the correlation coefficient at first; Then with Gaussian mixture noise, it studies the phenomena ofboth stochastic resonance and supra-threshold stochastic resonance in the maximizing network withthreshold devices based on Fisher information; Finally, we give the application of stochasticresonance in the encoding and decoding. This paper is divided into six parts:The first part describes the research background, the history of stochastic resonance study andthe current situation of research; then it introduces the theory of stochastic resonance briefly.The second part shows several kinds of common noise at first, and then displays somecommon stochastic resonance systems and introduces some main measures of stochastic resonance.The third part studies the phenomena of supra-threshold stochastic resonance in sum thresholdnetworks with the additive noise, multiplicative noise, or both additive and multiplicative noisebased on the correlation coefficient. It explores how the change of system’s threshold and thenumber of threshold units play the impact on the occurrence of stochastic resonance and on itsefficacy.The fourth part discusses stochastic resonance and supra-threshold stochastic resonance withGaussian mixture noise in the maximizing network with threshold devices. In this section, Fisherinformation acts as a measure of stochastic resonance.The fifth part discusses the simulation experiment about text transmission. By properly addinga small amount of noise to the system, letters can be received by comparing with the threshold. Thenumber of wrong letters may deduce gradually. However, if noise becomes stronger, the numbermay increase. And the phenomenon reflects the applicability of stochastic resonance in practice.The sixth part is a summary and outlook of this paper.