A holistic solution to pursuer-evader tracking in sensor networks

In this paper we devise a holistic solution to the pursuer-evader tracking problem taking into account the limitations of the wireless sensor networks (WSNs) as well as the dynamics of both the pursuer and evader. More specifically, we present an optimal strategy for the pursuer to capture the evader despite the delayed and imprecise information available at the pursuer-side. In order to minimize the communication overhead while ensuring capture, we provide an optimal evader-sampling scheme that adjusts the sampling frequency based on the strategies of the pursuer and evader, as well as the distance between the pursuer and evader. We support our adaptive sampling scheme with a just-in-time delivery protocol that publishes the evader's location updates directly to the pursuer, reducing the communication overhead of tracking even further. To further enhance the tracking reliability, we use a two-level design of fault tolerance: 1) a double position advertisement scheme to mask single message losses, and 2) a breadcrumbs-based backup scheme for stabilizing from desynchronization. Our simulation results show that the adaptive sampling scheme guides the pursuer to capture the evader effectively, and reduces the communication overhead significantly compared to fixed rate sampling. Our simulation results also show that our two-level fault-tolerance strategy ensures high capture rates even under consecutive message losses.