A data clustering approach to energy conservation in wireless sensor networks

This paper presents a new cluster-based power preservation scheme. Our clustering strategy for power saving is defined based on the “similarity of data” coming out from the sensors. The proposed clustering works in conjunction with our learning algorithm to obtain an optimum sleeping time of the sensors without disrupting the monitoring activity. The algorithm helps to effectively regulate the activation or deactivation of the sensor node’s radio transceiver, which in turn prolong the lifetime of the network. We have carried out several real-world experiments concerning power utilisation of wireless sensor devices in different scenarios and found that the proposed method leads to a significant power efficiency improvement with up to four times longer battery lifetime than other cases without such scheme.