MAMPS: Safe Multi-Agent Reinforcement Learning via Model Predictive Shielding

Abstract

Reinforcement learning is a promising approach to learning control policies for performing complex multi-agent robotics tasks. However, a policy learned in simulation often fails to guarantee even simple safety properties such as obstacle avoidance. To ensure safety, we propose multi-agent model predictive shielding (MAMPS), an algorithm that provably guarantees safety for an arbitrary learned policy. In particular, it operates by using the learned policy as often as possible, but instead uses a backup policy in cases where it cannot guarantee the safety of the learned policy. Using a multi-agent simulation environment, we show how MAMPS can achieve good performance while ensuring safety.

References in corpus (2)

• Safe Model-based Reinforcement Learning with Stability Guarantees
• Emergence of Grounded Compositional Language in Multi-Agent Populations

Cited by in corpus (5)

• Learning Safe Multi-Agent Control with Decentralized Neural Barrier Certificates
• Safe Multi-Agent Interaction through Robust Control Barrier Functions with Learned Uncertainties
• Multi-Agent Constrained Policy Optimisation
• Safe Deep Reinforcement Learning for Multi-Agent Systems with Continuous Action Spaces
• Safe Learning of Uncertain Environments