TARGET TRACKING BASED ON BASE STATION NODE USING PREDICTION METHOD AND CLUSTER STRUCTURE IN WIRELESS SENSOR NETWORKS

One of the most important and major challenging issues of wireless sensor networks is the tracking of mobile targets. The network continuously reports the spatial information of moving objects during specified periods to the base station. In this paper, by introducing new a protocol with two versions, of which, one of them is based on dynamic clustering with a focus on the base station, and the other is based on a predictive system for increasing the tracking accuracy of the objects movement and decreasing the energy consumption as well. In this paper, the task of clustering involves in determining the cluster heads, the number of cluster members, the selection of cluster members, and managing the activation of the nodes that is done by the base station. On the other hand, given that the base station is outside the field of wireless sensor networks and is connected to an unlimited power source. The second version of the proposed protocol is based on a predictive algorithm that it was inspired from the first proposed version in the role of the base station node by a prediction method. In this paper, three heuristic models are introduced to select the speed and direction in prediction models. They are instant, average and exponential-average models. These models can track the relevant targets more accurately and reduce the number of missing targets. The simulations are done in different scenarios in a custom developed tool. The results of simulation show a good performance of them in the network lifetime and target tracking applications.