A class of stochastic resonance systems for signal processing applications

The class of stochastic resonance systems based on M-level quantizer maps is developed. We derive expressions for the output invariant density and autocorrelation function of these maps when they are driven by a square wave in noise. These systems are shown to provide signal-to-noise ratio enhancement and robustness to noise characteristics. These properties render the quantizer maps potentially appealing for a wide range of signal processing applications such as interference suppression and robust communication. A framework for the analysis of more general discrete-time stochastic resonance systems is also presented, which is based on approximating these systems via quantizer maps