A Deep Reinforcement Learning Approach Using Asymmetric Self-Play for Robust Multirobot Flocking

摘要

Flocking control, as an essential approach for survivable navigation of multirobot systems, has been widely applied in fields, such as logistics, service delivery, and search and rescue. However, realistic environments are typically complex, dynamic, and even aggressive, posing considerable threats to the safety of flocking robots. In this article, based on deep reinforcement learning, an <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">A</i>symmetric <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">S</i>elf-play-empowered <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">F</i>locking <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">C</i>ontrol framework is proposed to address this concern. Specifically, the flocking robots are trained concurrently with learnable adversarial interferers to stimulate the intelligence of the flocking strategy. A two-stage self-play training paradigm is developed to improve the robustness and generalization of the model. Furthermore, an auxiliary training module regarding the learning of transition dynamics is designed, dramatically enhancing the adaptability to environmental uncertainties. Feature-level and agent-level attention are implemented for action and value generation, respectively. Both extensive comparative experiments and real-world deployment demonstrate the superiority and practicality of the proposed framework.