On robust k-hop clustering in ad-hoc cognitive radio networks

Cognitive radio networks (CRNs) enable cognitive users (CUs equipped with spectrum sensing) access the underutilized spectrum licensed to primary users (PUs) without causing unacceptable interference to the PUs' activities. On appearance of PUs, the available channel of CUs at different position may have different available channels which changes dynamically over time. Due to temporal and spatial variations of channel availability among CUs poses research challenges for ensuring connectivity and robustness of CRN. Reported works have shown effective use of clustering technique for network connectivity, cooperative spectrum sensing and a coordinated channel switching in CRN. Connectivity within the CRN is guaranteed as long as there is at least one channel available within each cluster and also between neighboring clusters. Most of the existing clustering schemes divide CRN into the least number of clusters based on the available channel common to the largest set of 1-hop neighbors. The drawbacks of these schemes are they do not provide robustness and require frequent re-clustering to maintain connectivity in CRN because of small number of common channel in each cluster. We have proposed a heuristic for k-hop clustering with objective to connect larger set of CUs in CRN. Our proposed algorithm converges in O(n.k), where n is the number of CUs and k is the number of hops. We have evaluated the performance of proposed scheme through simulation and observed that k-hop clustering algorithm achieves 40-50% more common channels as compared to other competitive approaches for k = 1 and improves robustness to 40%.