Energy Conservation Through Data Prediction in Wireless Sensor Networks

Abstract

Energy conservation is one of the main issue in wireless sensor networks (WSNs) which needs to be minimize. In WSNs, the sensor nodes have generally limited supply of energy due to limited size of the batteries. Since, the communication consumes maximum energy hence, by minimizing the number of communications in the network between sensor nodes and base station (BS) without compromising the quality of information be the logical way to conserve energy. This paper proposes a data prediction model that predict the next data within some limited error bound. Data prediction is one of the important techniques for data reduction. This data prediction model works on sensor nodes as well as on BS and makes the same predictions. The sensed data in WSN changes slowly with the time and follow some specific pattern. The variations in sensed data can be interpreted as an autoregressive model of order p i.e. AR(p). The order p is determined and it is observed that most of them are of order 1 only. AR(p) is one of the model of time series analysis which is a proven concept for predicting future data by understanding the past data.