Stochastic Resonance for Estimation of a Signal’s Bandwidth under Low SNR

Xiao Wang,Yong Gao
DOI: https://doi.org/10.1007/s10470-016-0809-y
IF: 1.321
2016-01-01
Analog Integrated Circuits and Signal Processing
Abstract:Traditional bandwidth estimation has low accuracy under low signal-to-noise ratio (SNR). To solve this problem, a new method of bandwidth estimation based on stochastic resonance is proposed. This method will process the signal twice by means of stochastic resonance, which improves the SNR while reducing fluctuations of the amplitude spectrum edge, and then calculating the local mean value function of the amplitude spectrum. A bandwidth can be estimated by using the starting and ending points where the amplitude value of the local mean function is greater than a certain threshold. The experimental simulations show that this method can accurately estimate the bandwidth of the phase shift key empty, quadrature amplitude modulation, and orthogonal frequency modulation at low SNRs. For example, the estimation accuracy could reach 90 % when the SNR is −10 dB. It is easy to conclude that the proposed method surpasses traditional bandwidth estimation method in terms of accuracy.
What problem does this paper attempt to address?