ARIMA-modeling based prediction mechanism in object tracking sensor networks

Target tracking is a typical and important application of wireless sensor networks which has gained much attention during recent years. Existing target tracking algorithms focus mainly on energy efficiency, maintenance of tracking accuracy and reducing the number of nodes involved in the tracking process with help of prediction mechanisms. Using prediction mechanisms with high accuracy and low computational complexity has an important role in reducing energy consumption and maintaining tracking accuracy. In this paper, we propose a novel algorithm called Auto-Regressive integrated Moving Average-based distributed predictive tracking (ARIMA-DPT) which presents a prediction model with high accuracy for prediction of target next location using ARIMA time series. Experimental results obtained by NS2 simulator improves the efficiency of proposed ARIMA-DPT protocol in terms of prediction accuracy and reducing energy consumption of network.