Traffic light optimization with low penetration rate vehicle trajectory data.

Area of Science:

• Urban planning and intelligent transportation systems.
• Traffic engineering and network optimization.
• Data science and applied mathematics.

Background:

• Traffic light optimization reduces urban congestion and energy consumption.
• High costs of vehicle detectors limit optimization of fixed-time traffic signals.
• Existing systems often rely on detectors, increasing costs and maintenance.

Purpose of the Study:

• To develop a large-scale traffic signal re-timing system using minimal vehicle trajectory data.
• To overcome the limitations of detector-based systems for traffic signal optimization.
• To provide a scalable and sustainable solution for urban traffic management.

Main Methods:

• Developed a probabilistic time-space diagram linking a stochastic point-queue model with vehicle trajectories.
• Reconstructed spatial-temporal traffic states by aggregating historical vehicle trajectory data.
• Applied optimization algorithms to update traffic signal parameters for improved efficiency.

Main Results:

• A citywide test in Birmingham, Michigan, demonstrated significant improvements.
• Reduced intersection delay by up to 20%.
• Decreased the number of vehicle stops by up to 30%.

Conclusions:

• The developed system offers a scalable, sustainable, and efficient approach to traffic light optimization.
• Eliminates the need for expensive vehicle detectors, making optimization accessible.
• Has the potential for widespread application to fixed-time signalized intersections globally.