Simplification of 3-level Quantizer Stochastic Resonance in Gaussian Mixture Modeled Non-Gaussian Noise

摘要

In non-Gaussian noise, 3-level quantizer stochastic resonance is a noticeable method for the purpose of signal detection. However, its performance depends on the optimal threshold which should be computed by the estimation of the ambient noise. In Gaussian Mixture modeled noise, the computation burden limits the use of this type stochastic resonance in several applications. In this paper, we considered the simplification of 3-level quantizer stochastic resonance. The comparison with Rao test is given. It shows that when the variance of noise is known, the difference between two types stochastic resonance is not prominent. Therefore, the simplification of 3-level quantizer is practical for detection.