| Stochastic resonance(SR)describes that the energy of noise may transfer to the signal when noise matches the input signal and the nonlinear system.It means that the noise is reduced and the signal energy is enhanced at the same time.Besides,the output signal-to-noise ratio(SNR)increases gradually and reaches the maximum point.As a result,noise can be considered as a favorable factor based on SR,which is of significance to signal processing in nonlinear systems.This paper studies the realization and application of SR in channel decoding.The main parts are as follows:This paper firstly studies the SR phenomenon and the performance of LDPC decoding improved by Gaussian noise in a threshold array system.Here bit error ratio(BER)is used to measure the decoding performance.The effects of the number of threshold units as well as the threshold value are discussed explicitly.We find that when signal is subthreshold,the BER declines first and then rises as noise intensity increases,which means SR exists.Especially,when the number of threshold units is large,SR is shown obviously.However,when signal is suprathreshold,noise does not promote the performance of decoding,which means SR does not occur.Then,this paper explores the performances of LDPC decoding promoted by Rayleigh noise and Rice noise in the threshold array system.Such two types of noise are common in image transmission.The simulation results show the differences between different noises.Similarly,the effects of the number of threshold units as well as the threshold value are discussed.And we confirm the existence of supra-threshold stochastic resonance(SSR)under Rice noise.Lastly,this paper focuses on the image transmission in the LDPC coding and decoding system.The effect of noise enhancing transmission is displayed based on peak signal-to-noise ratio(PSNR).And different noises which help improve image quality are further examined.The results demonstrate the applicability of noise in enhancing the performance of LDPC decoding. |
PDF Full Text Request