Cooperative sensing using probabilistic graphical models in Cognitive Radio networks

Abstract

We propose a new scheme for minimizing the probability of false alarm in cooperative CR (Cognitive Radio) networks by using graphical model theory. In order to improve the performance, we formulate a dynamical assignment for SUs (Secondary Users) to better detect PUs (Primary Users). A restriction in the assignment of SUs is adopted to reduce the complexity of optimal solution, and a sub-optimal message passing algorithm is presented based on the probabilistic graphical model.