A Cluster-Based Continuous Object Tracking Scheme in Wireless Sensor Networks

Abstract

Due to the severe resource constraints of sensor hardware, energy efficiency is a principal factor for detecting and tracking the movement of the large-scale phenomena such as wild fire, poison gas and hazardous bio-chemical material, denoted by continuous objects. In order to save the energy, the selective wakeup approach is effective way in the wireless sensor networks. However, most previous researches concentrated on individual objects such as intruders, tanks, and animals cannot be applied to the continuous object tracking because it is hard to expect that the diffusion of continuous object has the uniform velocity or acceleration. Recently, a prediction-based selective wakeup algorithm, denoted by PRECO is proposed. Nevertheless, this mechanism is still not acceptable. First, its prediction result is very inaccurate because a sensor node calculates next boundary line with only a few data while the continuous objects are pretty flexible and vicissitudinous. Second, its prediction tasks among current boundary nodes should be operated at the same time but it is a hard problem for the nodes to have synchronization in the wireless sensor networks. Therefore, we propose a novel prediction and selective wakeup scheme for energy efficient and accuracy continuous object tracking by using structured clustering. Hence, when objects are detected in some specific area, appointed areas where the target may move activate to keep guard without any complex calculations. Moreover, our scheme is asynchronous so that it is suitable for the sensor networks.