A rule learning approach to energy efficient clustering in wireless sensor networks

Physical clustering in wireless sensor networks results in the nomination of 'cluster heads'. The cluster head acts as a hub for the cluster. It is a specific node which has superior energy capabilities when compared with the other members of the same cluster. The nomination of cluster head is performed periodically or iteratively. This process is termed as re-clustering. Reclustering is energy-consuming due to the exchange/broadcast of numerous messages. Thus, this paper uses a rule-learning framework, ARTS (adaptive rule triggers on sensors), to prolong the intervals betweeen reclustering and thus, reduce the number of messages exchanged. The aim is to conserve the energy of cluster heads by using rules obtained from learning/analysis in clustering processes. To demonstrate, we have used the state-of-the-art clustering protocol HEED (hybrid energy efficient distributed clustering) due to its high energy-efficiency in selecting cluster heads. From our experiments using ARTS with HEED, we show that the re-clustering process in any physical clustering algorithm can be performed in a more energy-efficient manner.