Different weighted fusion scheme for Cooperative Spectrum Sensing using Normal Factor Graph

Abstract

Cooperative Spectrum Sensing (CSS) is a reliable detection approach that performs better over faded channel. In CSS, all Secondary Users (SU) have equal weight. We have studied performance of Centralized CSS (CCSS) using Normal Factor Graph (NFG) model with logical OR and Neyman-Pearson based Likelihood Ratio Testing (LRT) fusion rules. Here, we have proposed a novel Weighted CCSS (WCCSS) using NFG based probabilistic inference model. Sum Product algorithm (SPA) has been used for message passing and different weight assignment strategies have been used. The performance of WCCSS approach in Cognitive Radio Network (CRN) under different channels have been studied through simulation.