LAUSR

Stochastic resonance (SR) is widely used as a signal enhancement technique in recovering and enhancing periodic or aperiodic signals submerged in noise. However, system parameters and noise intensity tend to influence the SR performance. To achieve better resonance performance, several indices are often used to determine these parameters, including signal-to-noise, amplification factor, and cross-correlation coefficient. Nevertheless, for a linear frequency modulated (LFM) signal, such indices may no longer work and consequently make SR unable to recover the unknown LFM signal from raw signals. Thus, this limits the application of SR to some extent. To deal with this problem, we define here a new index to characterize the unknown LFM signal with the help of the fractional Fourier transform. Guided by this index, SR is thus able to recover the unknown LFM signal from the raw signal. In addition, a cloud model based genetic algorithm is used to achieve an adaptive SR in order to improve the effectiveness of signal processing.