POSE: Phase I: OpenLibSignal: Building a Distributed Open-Source Ecosystem for AI-Enabled Urban Systems Research

Researchers use a variety of software tools and datasets to study and simulate traffic in cities. These tools are vital for city planning and proactively managing traffic. The OpenLibSignal project aims to develop an open-source ecosystem of users and content developers that can improve connectivity among three popular open-source traffic management tools and datasets - CityFlow, LibSignal, and RL-Signal. Creating an integrated ecosystem around these tools will allow for more reliable and reproducible analysis of traffic patterns among different sites, which will improve the quality of urban systems research and assist policymakers. The ecosystem will also serve to improve the security and long-term sustainability of these resources. OpenLibSignal is a community-driven, open-source ecosystem for reproducible, low-cost, and scalable research in artificial intelligence (AI)-enabled urban simulation. Despite the rapid growth of open-source platforms for traffic management and urban simulation, current systems remain fragmented and difficult to sustain. Existing tools are often tied to a single simulator, lack standardized evaluation workflows, and provide only ad hoc support for community contributions, making reproducibility and comparability across sites inconsistent. Governance and sustainability mechanisms are also underdeveloped, leaving projects dependent on small core teams and vulnerable to single points of failure. To address these challenges, this project unifies three existing open-source artifacts — CityFlow (simulator), LibSignal (benchmarking algorithms), and RL-Signal (datasets/pipelines) — into a single, comparable workflow that supports traffic signal control, multi-modal mobility planning, and sustainability-focused policy evaluation. Phase I (1) establishes a distributed development workflow with rigorous testing and reproducibility checks; (2) designs and pilots a governance model with clear roles, processes, and release policies; and (3) grows a developer–user–educator community through targeted onboarding, tutorials, and events. By lowering barriers for municipalities, startups, and academic groups to prototype and evaluate urban system strategies before field deployment, OpenLibSignal advances reproducible AI research and contributes to national goals in transportation efficiency, environmental stewardship, and urban resilience. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria. NSF Award ID: 2550203 | Program: 01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Hua Wei | Institution: Arizona State University, SCOTTSDALE, AZ | Award Amount: $298,888 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2550203 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2550203.html