Age-Optimal UAV Trajectory Planning for Information Gathering with Energy Constraints

Many time-critical networks based on real-time information gathering by unmanned aerial vehicles (UAV) from a set of ground terminals. For those networks, minimization of the age of information (AoI), a metric proposed recently to measure the freshness of information, is of great importance. In this paper, we consider the problem of minimizing peak age, a metric of AoI, of a network composed of a mobile agent, a charging station and ground terminals. The agent's mobility is constrained by a graph $G$ and energy constraints including battery capacity and charging rate. Aiming at this problem, we study Markov Process and solve it by dimensional reduction. We work out the theoretical minimum peak age under the constraints and propose the least-charging-timed Metropolis-Hastings trajectory, a semi-randomized trajectory proved to be theoretical optimal. Furthermore, we propose a heuristic trajectory named least-visit-time-based trajectory for the case that visit times for each ground terminal are available for the agent.