Zhang, Zhiyao; Zhang, Yuhang; Quiñones-Grueiro, Marcos; Barbour, William; Biswas, Gautam; Work, Daniel. “1000DaySim: Open-source traffic simulation with real data over long time horizons.” Proceedings of the ACM/IEEE 16th International Conference on Cyber-Physical Systems, ICCPS 2025, held as part of the CPS-IoT Week 2025 (2025). https://doi.org/10.1145/3716550.3725151.
This poster introduces 1000DaySim, a new open-source traffic simulation tool that models a small network of three intersections using real-world data collected over nearly three years (1,096 days) in Nashville, Tennessee. The goal is to help researchers and engineers better understand and improve how traffic signals are controlled, especially using AI methods.
The project highlights several key contributions:
- It shows how to build a traffic simulation using real, diverse data sources from the field.
- It demonstrates the value of having long-term data when comparing and improving traffic signal timing plans.
- It provides easy-to-use scripts so others can quickly run their own simulations using the library.
This work has the potential to make AI-driven traffic signal systems more reliable by accounting for how traffic naturally changes over long periods. The next step for the team is to test and compare existing traffic control methods using this real-world simulation environment.
