Integrated movement-level traffic arrival prediction and signal control optimization using AVI data

This study introduces an adaptive signal control framework that integrates traffic arrival prediction and signal timing optimization for urban intersections, leveraging Automated Vehicle Identification (AVI) data to account for movement-level arrival information and adapt to dynamic traffic conditions. A novel movement-based vehicle arrival prediction module is first developed. The predicted arrivals then serve as critical inputs to a rolling-horizon signal timing model, which is formulated as an integer programming problem minimizing the weighted total delay and is solved through a dynamic programming algorithm. A distinctive feature of this optimization is the incorporation of travel time-dependent movement-aware weighting factors, which prioritize green time allocation for congested movements exhibiting prolonged historical travel times. Extensive simulation experiments conducted in SUMO demonstrate the efficacy of the proposed framework. The integrated framework's applicability is underscored by its enhanced adaptivity under shortened optimization horizons, highlighting its potential for real-time implementation in adaptive traffic signal control systems.