Suprathreshold Stochastic Resonance And Speech Signal Enhancement In Multilevel Threshold System

In communication systems,the effective information is often affected by noise,which leads to the damage of the output information.Stochastic resonance theory points out that a moderate amount of noise can be transformed into a positive factor to promote information transmission.In this paper,the theory of suprathreshold stochastic resonance is used to enhance speech signals in communication systems.Specific work contents are as follows:Firstly,take the correlation coefficient as a measure.The input is Gaussian distributed signal.we discuss the suprathreshold stochastic resonance under additive noise,multiplicative noise and their combined effects in a multithreshold nervous system.The experiment proved increasing the threshold units and threshold output numbers can enhance the effect of suprathreshold stochastic resonance.And the threshold interval is important to promote the occurrence of the stochastic resonance.Secondly,take the correlation coefficient as a measure.The input is Laplace distributed signal.In the above model,suprathreshold stochastic resonance under the action of additive Gaussian noise and four types of classical multiplicative noises is discussed.Among these four types of noises,the others all promote the stochastic resonance,except for Cauchy noise.Besides,a simulation is done,choosing an audio file as input.It turns some noise can improve the transmission effect of speech signal according to the stochastic resonance.Finally,select SNR as criterion.And the input is speech signal.When it is affected by additive Gaussian noise and multiplicative Gaussian noise,an adaptive adjustment algorithm is proposed.The iterative algorithm defines the optimal noise intensity,maximizes the stochastic resonance,and realizes the enhancement of the speech signal.