Modified Kalman Filtering Framework Based Real Time Target Tracking Against Environmental Dynamicity in Wireless Sensor Networks

One of the most widely used economical approaches to localization and tracking of mobile target with wireless sensor networks (WSNs), is the use of received signal strength indicators (RSSIs). In this paper, a modified kalman filtering based approach of real time tracking of single target moving in 2-D in WSN, is presented to deal with uncertainties in measurement noises and abrupt changes in target velocity. Two algorithms namely, RSSI + kalman filter (KF) and RSSI + unscented kalman filter (UKF) are proposed to refine estimates of the traditional RSSI based approach to obtain a smoothed target trajectory. The performance of the proposed algorithms is investigated against environmental dynamicity such as abrupt variations in target velocity, the limited set of RSSI measurements and the variation in the anchor density. The results confirmed that the proposed algorithms achieve better tracking accuracy and real time performance, irrespective of environmental dynamicity, compared to the traditional RSSI based algorithm.