5G-Enabled Multi-Agent Reinforcement Learning Framework for (CAV) Coordination | NEUTC Webinar скачать в хорошем качестве

5G-Enabled Multi-Agent Reinforcement Learning Framework for (CAV) Coordination | NEUTC Webinar

Join us this time for our free virtual NEUTC Webinar Series session, “5G-Enabled Safe & Robust Deep Multi-Agent Reinforcement Learning Framework for CAV Coordination,” presented by Fei Miao with the University of Connecticut. This webinar highlights a recent NEUTC project focused on coordinating connected and automated vehicles (CAVs) in mixed-traffic environments where uncertainty, noisy sensors, and imperfect observations are the norm. While reinforcement learning (RL) is widely used for decision-making, most existing safe RL methods assume accurate state information and define safety only in terms of expected outcomes, making it difficult to guarantee safety at every time step. To address this, the research team developed a safety-guaranteed hierarchical coordination and control framework called Safe-RMM, designed to ensure robust, multi-agent coordination under uncertainty. Results show the approach delivers a strong and efficient performance in challenging mixed-traffic scenarios. Check out our website to learn more about this project and others.