Correlation-based clustering in wireless sensor network for energy saving protocol

Abstract

Our research concentrates on the energy efficiency in Wireless Sensor Network. One approach is based on the characteristics of environment -- the correlation among sensed data of nodes in a region. The sensor nodes are clustered into highly correlated regions (HCRs) to take advantage of correlation between sensor nodes in order to save energy. However, the determination of HCR is very complex in calculation, thus causes difficulty in implementation. This paper proposes a correlation-based approach that evaluates the correlation between two data sets using a simple calculation and that guarantees the accuracy in correlated evaluation between data. This correlation-based method is proposed to cluster sensor nodes into HCRs. Because of highly correlated characteristics among sensed data of nodes in the same HCRs, some high correlated-nodes would be inactive for energy saving. Simulation results show that the network lifetime of proposed system is 1.75 times longer than that of the conventional protocol.