Efficient target tracking applications require active sensor nodes to track a cluster of moving targets. Clustering could lead to significant cost improvement as compared to tracking individual targets. This paper presents accurate clustering of targets for both coherent and incoherent movement patterns. We propose a novel clustering algorithm that utilises an implicit dynamic time frame to assess the relational history of targets in creating a weighted graph of connected components. The proposed algorithm employs key features of localisation algorithms in target tracking, namely, estimated current and predicted locations to determine the relational directions and distances of moving targets. Our simulation results show a significant improvement on the clustering accuracy and computation time by dynamically adjusting the history-window size and predicting the relationships among targets.
Efficient target tracking applications use active sensor nodes collaboratively to track multiple moving targets by balancing the trade-off between the quality of tracking and network's lifetime. In this paper, we propose a low-energy dynamic sensor selection (LEDS) scheme to track multiple targets by estimating energy consumption of sensors and information utility projection of the targets on sensors to calculate the eagerness in tracking. Eagerness represents the eligibility of a sensor node to be selected for tracking, considering relative profiles of other sensors and location of all the targets in its vicinity. LEDS enables an even distribution of energy consumption among the nodes to prolong their remaining energies. Our results show that the proposed scheme can significantly improve the network lifetime over the existing methods while maintaining the high tracking accuracy in congested areas where multiple concurrent targets overlap.