Cooperative and Distributed Reinforcement Learning of Drones for Field Coverage

This paper proposed a distributed Multi-Agent Reinforcement Learning (MARL) algorithm for a team of Unmanned Aerial Vehicles (UAVs) that can learn to cooperate to provide a full coverage of an unknown field of interest while minimizing the overlapping sections among their field of views. Two challenges in MARL for such a system are discussed in the paper: firstly, the complex dynamic of the joint-actions of the UAV team, that will be solved using gametheoretic correlated equilibrium, and secondly, the challenge in huge dimensional state space will be tackled with an efficient space-reduced representation of the value function table. We also provided our experimental results in details with both simulation and physical implementation to show that the UAVs can successfully learn to accomplish the task without the need of a mathematical model. This paper can serve as an efficient framework for using MARL to enable UAV team to work in an environment where its model is unavailable. In the future, we are interested in accounting for the stochastic aspect of the problem and technical aspects of a real-world deployment, where uncertainties, such as wind and other dynamics of the environment presents.