Eigenvector centrality based cluster size control in randomly deployed wireless sensor networks

• We propose a new Eigenvector Based Cluster Size Control method Ev-CSC.
• Ev-CSC utilizes PageRank and α cut set to address uneven energy consumption.
• WSN is modeled as fuzzy graph to address dynamic and uncertain characteristics.
• Ev-CSC is independent of node density, shape of deployment region and sink position.
• Simulation results indicate improved network lifetime and energy efficiency.

Cluster size control plays a significant role in balancing energy consumption and mitigating hot spot problem in wireless sensor networks. Cluster size control is necessary as clusters having higher member nodes consume significantly higher amount of energy while low member nodes in a cluster lead to under utilization of channel capacity. Further, cluster-heads located near to sink have to perform additional function of relaying data of other nodes. All these factors are responsible for hotspots or energy holes creation which in turn affect the network lifetime. In this paper, we propose a heuristic approach based upon Eigenvector centrality for cluster size control which we have named as Ev-CSC. While existing methods in literature consider distance of a cluster-head from sink, layered architecture, uniform deployment of nodes, prefixed sink location as a precondition, our proposed method does not have these constraints and is applicable to any kind of deployment, traffic pattern and node types. We have applied Ev-CSC on equal clustering methods and have also compared the same with state-of-the-art cluster size control methods. The experimental results demonstrate that our proposed method enhances the performance of respective equal clustering methods and performs better as compared to cluster size control methods.