Analysis of adaptive clustering algorithms in wireless sensor networks

The paper presents the criteria that affect the performance of network lifetime in Wireless Sensor Networks. In sensor networks the nodes can be organized into clusters for one hop energy-efficient data routing. This work focuses on adaptive clustering of sensor nodes where the role of cluster head changes among nodes. The nodes take turn to become cluster head based on one or several parameters such as random selection, nodes residual energy, nodes ID, node proximity to other nodes or nodes with different energy levels (advanced nodes). Ideally, the number of elected cluster heads must match the optimal number of cluster heads which is determined a priori or via simulation for energy-efficient routing. Achieving the optimal number of cluster heads is crucial as this will ensure the minimum energy consumption for data transmission to base station. This paper presents simulation results of existing adaptive clustering algorithms. The simulation results show how the election criteria for cluster heads election such as node residual energy, node proximity to other nodes, random election and nodes with different energy level affect the number of cluster heads elected, their distribution in the network and the network lifetime. Simulation results are provided to show the comparative effectiveness of different clustering algorithm on network lifetime, cluster head selection and their distribution in the network. Modifications on the cluster head election criteria are done that combined the node residual energy and nodes proximity to other nodes and simulation shows that the network lifetime is improved.