Autonomous Mobility on Demand

Autonomous Mobility-on-Demand (AMoD) systems aim to optimize the efficient and equitable allocation of self-driving vehicles to meet dynamic transportation demands. Current research heavily focuses on developing robust and scalable control algorithms, employing techniques like multi-agent reinforcement learning (MARL), graph reinforcement learning, and large multimodal models to address challenges in real-time dispatching, fleet rebalancing, and charging optimization for electric vehicles. These advancements hold significant promise for improving urban transportation efficiency, reducing congestion and emissions, and enhancing accessibility, particularly through the development of data-driven and robust solutions that account for uncertainties in demand and supply.