CAC: Context adaptive clustering for efficient data aggregation in wireless sensor networks

Wireless sensor networks are characterized by the widely distributed sensor nodes which transmit sensed data to the base station cooperatively. However, due to the spatial correlation between sensor observations, it is not necessary for every node to transmit its data. There are already some papers on how to do clustering and data aggregation in-network, however, no one considers about the data distribution with respect to the environment. In this paper a context adaptive clustering mechanism is proposed, which tries to form clusters of sensors with similar output data within the bound of a given tolerance parameter. With similar data inside a cluster, it is possible for the cluster header to use a simple technique for data aggregation without introducing large errors, thus can reduce energy consumption and prolong the sensor lifetime. The algorithm proposed is very simple, transparent, localized and does not need any central authority to monitor or supervise it.