An agent-based modeling framework for the multi-UAV rendezvous recharging problem

In this work, we aim to model the multi-UAV rendezvous recharging problem, which consists of energy -limited aerial vehicles that rendezvous with a mobile or fixed charging station. The motivation for such a problem is to tackle persistent surveillance missions, where the modeling of related problems often rely on heavy mathematical formulations, such as mixed integer linear programming (MILP). The major drawback of such approaches is great difficulty in capturing constraints and adjusting the model for changes. Additionally, MILP solvers are not guaranteed to yield a feasible solution in a timely manner. As a result, we chose to create an agent-based model (ABM). To the best of our knowledge, the presented problem has not been modeled before using ABM. Additionally, we sought to use a custom framework incorporating Behavior Trees (BTs) and Hierarchical Finite State Machines (HFSMs), two commonly used tools in the video game and robotics industry. We verify the model's correctness through numerical simulations and show that it is highly modular and extensible.(c) 2023 Elsevier B.V. All rights reserved.