Distributed and Collaborative Primary Signal Feature Estimation for Cognitive Radios under Communication Constraints

Collaborative algorithms are needed to improve the reliability of spectrum sensing in a network of cognitive radios (CRs). This work studies a consensus based approach to sharing spectral measurements between a multihop network of CRs. Specifically, the impact of link errors are incorporated in determining the convergence behavior of consensus based spectrum sensing. Results show that as the number of message exchanges increases, the convergence time and the deviation of the convergence value increase. Hierarchical consensus, a modification to the original consensus algorithm, is proposed to reduce the number of message exchanges while still obtaining the collaborative gains of shared spectrum sensing.