Transmission Of Voice Signals Over The Threshold Of Stochastic Resonance

Along with the advent of an aging ring society problem、abundance of human material life level, the whole human sensory function and nutrition metabolism is reducing in different degrees. Improving signal detection technology and the human sensory function is becoming the urgent requirement in the real world. Scholars trying to reduce or suppress background noise to improve the performance of the system for a long time, found out the phenomenon that the noise in a certain conditions in a nonlinear system can help the signal transmission, which can called the phenomenon of "counter-intuitive",this phenomenon is named stochastic resonance. Stochastic resonance enhancing the detection performance of signal in noise background provides a new thought about signal detection technology, also makes a new processing technology about the overall decline of the sensory function in human beings.This study select cross-correlation coefficient as the evaluation index, analyzing the super threshold phenomenon stochastic resonance when the speech signal is transmitted by the dynamic sensory neurons group attacking by uniform distribution and exponential distribution noise and the change of speech signals short time domain features. First, using the uniform distribution and exponential distribution noise simulate the group internal noise in sensory nerve cells, make multiple neural model by parallel connection to simulate the sensory neurons group, and select two different types of voice signal as a system input signals. Second, chose cross-correlation coefficient as the evaluation index, quantitatively observe the different number of sensory neurons system output speech signal and the effect of sensory neurons’short time domain features of speech signals change with the two different kinds of noise in this case, and use short time Fourier spectrum analysis sensory neurons system’s output speech signal fundamental frequency. Finally, according to the actual data experiment, it has proved that super threshold stochastic resonance phenomenon in sensory neurons system and the enhancing impact on the short time domain features of speech signals under the internal sensory neurons noise environment. The principle of super threshold stochastic resonance and voice signals in the nervous system provides effective basis for the actual transmission problem, and speech synthesis, speech recognition technology in nerve cells of the group study also provides valuable reference.