The most challenging aspect of WSN is that they are energy resource-constrained and that energy cannot be replenished. Clustering methods in WSN lead the sensor nodes to be organized into small disjoint groups, where each cluster has a coordinator referred as cluster head (CH). Maintaining the clusters is a challenging task. To choose CH as a node, it is necessary to define its eligibility. That is calculated based on local information of the nodes’ current situations such as its residual energy. The eligibility of the selected CHs reduces as the sensor nodes are consuming energy for transferring data. However, there is possibility that the CHs may fail to function incorrectly due to various reasons such as power instability. During the failure, the CHs are unable to collect and transfer data correctly. This affects the performance of the WSN. Early detection of failure will reduce the data loss and provide possible minimal recovery efforts. This paper proposes a selfconfigurable clustering mechanism to detect the disordered CHs and replace them with other nodes.

Keywords: Residual Energy, Node Centrality, Local Distance, Cluster Heads