
The author proposes DeepSafeMPC, a method that combines deep learning with model predictive control to enhance safety in multi-agent environments. By leveraging a robust dynamics predictor, the approach ensures accurate future state predictions and effective decision-making within safe boundaries.


coremsg

deepsafempc-integrating-deep-learning-with-model-predictive-control-for-safe-multi-agent-reinforcement-learning


DeepSafeMPC: Integrating Deep Learning with Model Predictive Control for Safe Multi-Agent Reinforcement Learning
