Extending Max-Pressure Control for Traffic Network Operations

Traffic signal timing-ways to distribute right of way at signalized intersections-can significantly affect vehicle flows and create delays in urban cities. This project develops a new adaptive method for signal timing (max-pressure control) through mathematical proof and experimental results. Max-pressure control uses vehicle queue length information to optimize signal timing. This research addresses key fundamental limitations with existing work that discourage implementation and limit the practical benefits. This project also includes a pilot deployment study by partnering with Hennepin County in Minnesota to evaluate potential benefits on their intersections. Finally, it will extend the mathematics to suggest alternative route choices to minimize total congestion, and maximizing passenger flow in ridesharing services such as Uber and Lyft. Mathematical knowledge gained from this project will be directly taught through graduate school courses and Ph.D. student mentoring. To improve public awareness of the studied problems, outreach activities will be conducted to K-12 students.

This project will introduce realistic traffic flow assumptions into existing store-and-forward queueing models, namely high-density restrictions on flow, queue spillback, and first-in-first-out behavior. Since max-pressure signal timing normally relies on sensors to estimate queue lengths, it will investigate alternative data sources to calculate queue pressure, i.e. road travel delays. During these tasks, this project will search for new Lyapunov functions and modify the max-pressure control to retain the maximum-stability results. Simulation results will be used to study the effects on congestion when users change their routes in response to new signal timings. The project will further investigate actual benefits by partnering with Hennepin County in Minnesota to perform microsimulation and a pilot deployment on their intersections. Finally, the project will extend the throughput-maximizing approach to system optimal dynamic traffic assignment and passenger service in mobility-on-demand by using queueing models for those problems. Essentially, this research will bring the attractive analytical properties of max-pressure control into a more accurate model to move towards effective implementation in urban cities to realize the benefits in practice. The intellectual challenges of successful Lyapunov-based proofs of maximum-stability are significant. In addition, applications of the max-pressure approach to other transportation engineering problems with similar modeling goals will broaden the impacts across the transportation field.

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.