A Time-Frequency Analysis Method Using Stochastic Resonance

Zhao Zhijin,Yao,Xu Chunyun

DOI: https://doi.org/10.1109/cisp.2008.243

2008-01-01

Abstract:Time-frequency analysis has been widely used in non-stationary signal analysis, and the reassignment method can improve the performance of the time-frequency analysis. But they can not analyze signals at low signal-to-noise ratio. The stochastic resonance system can increase signal-to-noise ratio. Using this feature this paper proposes a time-frequency analysis method applying stochastic resonance as a pre-processing system. It means that a signal is pre-processed by stochastic resonance system first, and then is analyzed using time-frequency distribution and reassignment method. The simulation results show that the proposed method can analyze signals very well under low signal-to-noise ratio. When signal-to-noise ratio is -11dB, the reassignment spectrogram using stochastic resonance provides a higher time-frequency concentrate and improves the readability of spectrogram; the reassignment scalogram using stochastic resonance has the good performance.