Diffusion Distance-Based Predictive Tracking for Continuous Objects in Industrial Wireless Sensor Networks

In an industrial production process, the leakage of continuous objects poses a serious threat to production safety. In this paper, a diffusion distance-based predictive tracking algorithm is proposed for industrial wireless sensor networks (IWSNs), aiming to timely track the boundary of a continuous object after the occurrence of a leak. Based on the assumption that the motion of the continuous object follows an appropriate diffusion model, sensor nodes are able to capture environmental parameters for establishing the mathematical expression of the model locally. Through building up the relation of diffusion radius with time, each node predicts diffusion scope of the continuous object at different times and makes a judgment about whether it is suitable to be a boundary node. Moreover, to achieve high energy-efficiency, a sleep/wake cycle is introduced to involve a small number of nodes in the process of tracking, while the rest of nodes stay idle until an object approaches. Finally, a cluster-based competitive mechanism is proposed for reporting the location of boundary nodes. Simulation results demonstrated that our proposal is able to track the diffusion of continuous objects with high energy-efficiency.