Stochastic Resonance in Signal Detection and Human Perception

摘要

Stochastic Resonance (SR) is a nonlinear phenomenon first reported in terms of a nonlinear dynamic effect. The important question of what type of noise to be added has until recently evaded a solution. This issue is addressed directly and a fundamental theoretical framework is developed leading to a determination of the optimal. additive SR noise to achieve maximum probability of detection subject to the constraint that the probability of false alarm is not increased. Chapters 2 and 3 provide alternative analytical framework presentations for SR application to detection leading to an optimization solution. Chapter 4 discusses probability of error reduction with implications for communications theory. Subsequent chapters address applications of the analytical theory to suboptimal detectors such as nonparametric detectors (Chapter 5), image enhancement (Chapter 6), and distributed sensor fusion (Chapter 7). Chapter 8 provides a novel consideration to an alternative decision statistic transformation methodology to recover optimal performance for a suboptimal detector. Finally, Chapter 9 resents recommendations and future considerations.