Application oriented connected dominating set-based cluster formation in wireless sensor networks

Clustering is a fundamental mechanism used in the design of Wireless Sensor Network (WSN) protocols. The performance of WSNs can be improved by selecting the most suitable nodes to form a stable backbone structure with guaranteed network coverage. This paper proposes a base station-controlled centralized algorithm for static sensor networks and a distributed, weighted algorithm for dynamic sensor networks. The solutions are based on a (k,r)-Connected Dominating Set, which is suitable for cluster-based hierarchical routing. The clusterhead redundancy parameter k improves reliability, the multi-hop parameter r addresses the scalability issue and the combined weight metric improves the network lifespan and reduces the number of re-affiliations. To create a stable and efficient backbone structure, the backbone sensor nodes are selected based on quality, which is a function of the residual battery power, node degree, transmission range, and mobility of the sensor nodes. Simulation experiments are conducted to evaluate the performance of both the algorithms in terms of the number of elements in the backbone structure, re-affiliation frequency, load balancing, network lifespan, and the power dissipation. The results establish the potential of these algorithms for use in WSNs.