Distributed spectrum sensing using belief propagation framework

In this paper, cognitive secondary users (SUs) are taken into account to cooperate for messages information transmission without a data center using Belief Propagation - based framework (BP). We also exploit generalized belief propagation (GBP) to simulate individual marginal distribution of SUs in random topology networks with loops. Simulation results indicate that BP gives a substantial gain of around 30% to 60% over the single node detector (no cooperation sensing) especially in low false alarm rate, GBP outperforms the ordinary BP algorithm at the cost of time complexity and computational agility. The simulation results and illustrations verify these points, therefore, cooperation is necessary to sense spectrum in cognitive radio.