Automatic Generation of Traffic Signal Timing Plan
Henry Liu, Associate Professor, Civil, Environmental and Geo-Engineering
Project Summary:It is commonly understood that traffic delay increases 3 to 5 percent per year simply because signal timing plans are not kept up to date. Signal retiming, however, is only conducted once every 3 to 5 years because it involves largely manual processes. For many agencies with limited resources, it would be desirable to replace the labor-intensive signal retiming process with algorithms for automated or semi-automated updating of the signal parameter settings. Previous attempts have largely failed due to the unavailability of field performance data and the need for calibrating parameters of embedded traffic models. However, over the past decade or so, significant advances in data collection and performance measurement technology (for example, the SMART Signal system developed at the University of Minnesota) warrant a fresh look at automated timing plan generation. The goal of this research is to develop methods to automate coordinated timing plan generation on signalized arterials. This project aims to extend SMART Signal's capability for automatically diagnosing operational problems and fine-tuning signal control parameters. Automatic diagnosis and optimization tools will be developed based on the large volume of traffic signal performance data collected by the SMART Signal system. The data-driven algorithms and tools will extend the life of an existing signal timing plan by fine-tuning the parameters and will significantly reduce the frequency and expense of manual work for signal retiming, thereby bringing revolutionary changes to the practice of signal control systems.
- Start date: 11/2012
- Project Status: Completed
- Research Area: Transportation Safety and Traffic Flow
- Topics: Data and modeling, Traffic operations