Research

- Developing control strategies and algorithms that can effectively control intersection traffic flow under saturated conditions, which are difficult to explain with traffic detection data alone, by developing an AI model that imitates the empirical knowledge of experts on the target intersection - Based on this, developing V2X wireless communication-based vehicle control strategies for the future autonomous driving era - Ultimately aiming to verify the feasibility of the imitation learning-based AI model approach to traffic flow control

1. Development of imitation learning-based intersection control algorithms - stable control even under traffic congestion, applicable to smart city and autonomous driving traffic management 2. Establishment of a V2X-based simulation/testing platform - creation of a SUMO-Unity integrated simulation environment, verification of integrated management of autonomous driving and infrastructure 3. Verification and improvement of traffic control performance - optimization based on mobility/stability indicators, improvement of traffic policies/algorithms and establishment of a foundation for industrial application