Versatile Robust Clustering of Ad Hoc Cognitive Radio Network

Cluster structure in cognitive radio networks facilitates cooperative spectrum sensing, routing and other functionalities. The unlicensed channels, which are available for every member of a group of cognitive radio users, consolidate the group into a cluster, and the availability of unlicensed channels decides the robustness of that cluster against the licensed users’ influence. This paper analyses the problem that how to form robust clusters in cognitive radio network, so that more cognitive radio users can get benefits from cluster structure even when the primary users’ operation are intense. We provide a formal description of robust clustering problem, prove it to be NP-hard and propose a centralized solution, besides, a distributed solution is proposed to suit the dynamics in the ad hoc cognitive radio network. Congestion game model is adopted to analyze the process of cluster formation, which not only contributes designing the distributed clustering scheme directly, but also provides the guarantee of convergence into Nash Equilibrium and convergence speed. Our proposed clustering solution is versatile to fulfill some other requirements such as faster convergence and cluster size control. The proposed distributed clustering scheme outperforms the related work in terms of cluster robustness, convergence speed and overhead. The extensive simulation supports our claims.

💡 Research Summary

The paper addresses the need for robust clustering in cognitive radio networks (CRNs), where primary user (PU) activity can disrupt secondary users’ communications. Existing clustering schemes either ignore robustness against PU activity or focus solely on metrics such as sensing accuracy or energy efficiency. The authors define robustness as the ability of a cluster to survive increasing PU activity, which they quantify by the number of common licensed channels (CCs) shared by all members of a cluster. More CCs imply a longer expected lifetime before all channels become occupied by PUs.

The system model assumes a set N of secondary users (SUs) and a set K of licensed channels. Each SU i independently senses the spectrum and obtains a set K_i of currently available channels. A unit‑disk model defines one‑hop neighborhoods: two SUs are neighbors if they are within transmission range r and their available channel sets intersect. A dedicated control channel is used for exchanging K_i information.

A cluster C is defined as a set of SUs that share the same CC set K(C)=∩_{i∈C}K_i and have a designated cluster head. The problem is to partition all SUs into disjoint clusters such that (1) each cluster size lies within a desired interval