The location information of a mobile node is an essential parameter for vehicle monitoring and other location-based services (Junglas and Watson, Commun ACM 51(3):65---69, 2008). The conventional methods used for target tracking, which are applicable for vehicular networks, are either Global Positioning System-based, time of propagation-based or signal strength-based. All of these methods have their own limitations such as additional hardware requirement, power consumption, lack of accuracy, and environmental dependencies. Besides, traditional tracking algorithms do not consider the presence of misbehaving nodes in the network. In this paper, we study target tracking in vehicular mobile WiMAX network environments. We present the proposed Selfishness-Aware Target Tracking (SATT) algorithm. SATT uses time difference of arrival based measurement data when the target is in line-of-sight (LOS) with more than three base stations (BSs). When no more than three LOS links between the target and the BSs are available, then the cluster-head, which serves the target at that instant, activates the three most promising mobile nodes for collecting location information of the target. We use the Stochastic Learning Weak Estimator (Oommen and Rueda, Pattern Recogn 39(3):328---341, 2006) method for keeping track of the misbehaving nodes in the network. The volunteer nodes are selected for target tracking based on this information. Unscented Kalman Filter is used for estimation of the position and velocity of the target. The simulation results show that the SATT algorithm increases the accuracy in tracking information up to 70 % in comparison to the other methods and the algorithm is able to achieve 55---95 % cooperation depending on the degree of misbehavior in the network.
Read full abstract