Noise tuning in stochastic systems

Abstract

Noise-added systems present interesting features in many application fields and terms such as stochastic resonance and dithering are becoming more common. The main problem in dealing with noise-added systems is the identification of the optimal value of the noise variance as a function of system parameter fluctuations. In a previous work the authors introduced a theoretical relationship attesting the dependence of the noise variance on some forcing signal parameters. This relationship was used to preserve optimal system performance as the input signal characteristics change. In this paper both a numerical and a hardware solution for the implementation of the above-mentioned relationship are proposed. More specifically, both a virtual instrument and an analogue device implementing noise variance control are proposed.