Deep Reinforcement Learning-Based Approach for a Single Vehicle Persistent Surveillance Problem with Fuel Constraints

This article presents a deep reinforcement learning-based approach to tacklea persistent surveillance mission requiring a single unmanned aerial vehicleinitially stationed at a depot with fuel or time-of-flight constraints torepeatedly visit a set of targets with equal priority. Owing to the vehicle'sfuel or time-of-flight constraints, the vehicle must be regularly refueled, orits battery must be recharged at the depot. The objective of the problem is todetermine an optimal sequence of visits to the targets that minimizes themaximum time elapsed between successive visits to any target while ensuringthat the vehicle never runs out of fuel or charge. We present a deepreinforcement learning algorithm to solve this problem and present the resultsof numerical experiments that corroborate the effectiveness of this approach incomparison with common-sense greedy heuristics.