Abstract
In this paper we illustrate observations collected during the 2022 Aquaticus competition. In particular, we illustrate our experiences in terms of classical controller designs for capture the flag games. The Aquaticus competition is a capture the flag multi-player game where two teams of (autonomous) robots compete with each other. We illustrate ideas of a simple controller that combines ideas from optimal control, model predictive control and path planning using Dubins curves. The corresponding controller outperformed the other controllers in the 2022 Aquaticus competition. The results are meant as a stepping stone to combine classical controller designs with reinforcement learning. While reinforcement learning based control laws might be well suited for competitive controller design in future competitions, based on the authors experience, a good understanding of the problem seems to be inevitable to initiate a reinforcement learning controller design.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have