Adaptive stochastic resonance for unknown and variable input signals

Abstract

All sensors have a threshold, defined by the smallest signal amplitude that can be detected. The detection of sub-threshold signals, however, is possible by using the principle of stochastic resonance, where noise is added to the input signal so that it randomly exceeds the sensor threshold. The choice of an optimal noise level that maximizes the mutual information between sensor input and output, however, requires knowledge of the input signal, which is not available in most practical applications. Here we demonstrate that the autocorrelation of the sensor output alone is sufficient to find this optimal noise level. Furthermore, we demonstrate numerically and analytically the equivalence of the traditional mutual information approach and our autocorrelation approach for a range of model systems. We furthermore show how the level of added noise can be continuously adapted even to highly variable, unknown input signals via a feedback loop. Finally, we present evidence that adaptive stochastic resonance based on the autocorrelation of the sensor output may be a fundamental principle in neuronal systems.

The Stochastic Resonance model of auditory perception: A unified explanation of tinnitus development, Zwicker tone illusion, and residual inhibition

“…In self-adaptive signal detection systems exploiting SR, the optimum intensity of the noise is continuously adjusted so that information transmission is maximized, even if the characteristics and statistics of the input signal change (Figure 2). For this processing principle, the term adaptive SR has been coined (Mitaim & Kosko 1998Wenning & Obermayer 2003;Krauss et al, 2017).…”

Intrinsic noise improves speech recognition in a computational model of the auditory pathway

“…The term adaptive SR was coined for this processing principle [18][19][20]. In a previous study we demonstrated that the auto-correlation of the sensor output, a quantity always accessible and easy to analyze by neural networks, can be used to quantify and hence maximize information transmission even for unknown and variable input signals [21].…”

Effect of scale-varying fractional-order stochastic resonance by simulation and its application in bearing diagnosis

“…11,13,14,18,29,30 However, they are unavailable when the target frequency is unknown. As to this problem, some criterions are proposed such as approximate entropy, 12 MPSK index, 17 mutual information, 31 and weighted power spectrum kurtosis. 32 This study focuses on the relationship between the parameters of SFrSR and its output.…”