Statistical Data Aggregation Protocol Based on Data Correlation in Wireless Sensor Networks

Data aggregation schemes for wireless sensor networks have been used as mechanisms to generally optimize energy consumption. The schemes are used to reduce the total amount of sensor readings by means of performing a variety of aggregate operations and collaborating with sensor nodes in data transmission. In addition, it is known that sensor readings obtained from sensor fields have high spatial and temporal characteristics in nature. In this paper, we propose a novel energy-efficient data aggregation protocol that makes it possible to minimize energy consumption and to maximize communication efficiency with data correlation. The proposed protocol is comprised of two methods which satisfy the quality of application requirements when a user requests periodic data from a wireless sensor network. The first is a selective transmission method using a tolerable threshold of data loss rate. The other is a data aggregation clustering method which satisfies data accuracy of application requirement with linear data prediction. In addition, data delay can be minimized by allocating a weight to the size of each cluster and scheduling time-outs for the response data. These differentiated strategies will result in the longer network lifetime as energy consumption is evenly balanced throughout the wireless sensor network.