Secure spectrum sensing in relay-based cognitive radio networks

A novel cooperative spectrum sensing algorithm intended for robust sensing in the presence of Byzantine attacks was formulated for relay-based cognitive radio networks, with the computation distributed among the relay nodes. The development of this algorithm improves the viability of the application of decentralised CRNs to sensing, which in turn increases the number of use cases for the technology. The proposed algorithm, which performs probabilistic inference using belief propagation, learns from historical sensing results to provide an estimate of the primary user (PU) state. The algorithm was found to reduce the impact of malicious users significantly in comparison to the majority decision rule even in cases where the majority of users were malicious. Furthermore, the algorithm's PU state estimations converged quickly. Characteristic of a distributed algorithm, the performance of the algorithm was sensitive to the measurement quality of the relay nodes responsible for computation.