A Cascaded Approach to Optimal Aircraft Trajectories for Persistent Data Ferrying

Unmanned aircraft are ideal vehicles for collecting and transferring data for sparse sensor networks, acting as data ferries for the networks. This paper investigates the optimality of periodic trajectories of an unmanned aircraft ferrying data between two stationary sensor nodes. The aircraft has a single half-duplex radio channel for wireless communication, and must allocate the channel along its trajectory to maximize data transfer between the nodes. The general data ferrying problem is known to be NP-Hard. This work takes advantage of the problem structure to cascade the solution between trajectory optimization and bandwidth optimization. We prove that this decomposition retains optimality, and further derive necessary conditions for trajectories to be optimal. In particular, the ferry must deliver exactly as much data as it collects over the periodic trajectory. Insights from these conditions lead to well-defined policies on bandwidth allocation. Our cascaded approach thus significantly reduces dimension of the persistent data ferrying problem.