An Energy Efficient Hybrid Clustering Algorithm Combined with Prediction Method for Target Tracking in Wireless Sensor Networks

Target tracking in WSN has attracted a great attention duo to its growing application potential in different fields. One of the main problems for target tracking in WSN is to maximize network lifetime by reducing energy consumption as well as guaranteeing the target tracking quality at a certain level. Among different target tracking schemes, hybrid clustering resolves the boundary problem and guarantee the target tracking quality because the static cluster and the on-demand dynamic cluster take turns each other to track the target in hybrid clustering scheme. However, huge amount of energy can be consumed due to the frequent formation and dismiss of redundant dynamic clusters when the target zigzags between a static cluster and a dynamic cluster or when the movement of target makes overlapped dynamic clusters to be formed continuously. In order to resolve this kind of problems, in this paper, a hybrid clustering algorithm combined with prediction method is proposed so that energy consumption due to the overforming of dynamic clusters could be reduced and the target tracking quality could be guaranteed simultaneously. Furthermore, a scheme to adjust the size of predicted clusters and the length of target interval time, according to prediction error and target speed, is applied to guarantee the target tracking quality of the prediction-based clustering algorithm. The results of extensive simulation experiment show that the proposed scheme can guarantee the target tracking quality and extend network lifetime significantly although a huge amount of energy is consumed due to overforming and overdismissing dynamic clusters.