A Pre-Selected Clustering Protocol Based on Spatial Correlation

Wireless sensor networks are mainly characterized by dense deployment of sensor nodes which collectively transmit information about sensed events to the sink. Due to the high density of nodes, spatially redundant data is generated. In this paper, a pre-selected clustering protocol is proposed to exploit the spatial correlation. In each round, the protocol dynamically divides the entire sensing region into several correlated sub- regions and selects one node as the pre-selected node (representative node) of the sub-region. Only the pre-selected node in each sub-region transfers the sensed data. The pre-selected nodes execute dynamic clustering protocol later in the round. Simulations show that the new protocol can decrease the energy consumption of the nodes and improve the system lifetime under a given distortion constraint.