Home About ATSIP Program Awards Venue Sponsors/Exhibitors

Radar-Based Vehicle Detection to Continuously Monitor Volume and Speed at Signalized Intersections


Abstract: Continuously monitoring speed and volumes at signalized intersections is a difficult challenge that often requires detailed loop detector installations and the use of high-resolution data collection that is only possible with newer signal controllers. Moreover, even when high-resolution data collection is possible, the configuration of the intersection lanes makes if difficult to understand the volume by movements due to the presence of shared movement lanes and the inability of loop detectors to break vehicle detection data by movement. Therefore, even with the advances in technology, the standard practice for conducting traffic studies still involves the use of some form of manual labor in the process. As a result, obtaining datasets that are continuously updated and that allow understanding the historical traffic conditions of an intersection approach at a given point in time is a difficult task. Fortunately, even when the approach used to monitor vehicle presence at intersections has not changed in decades, the detection technology used to report the presence of vehicles in a legacy format has continued to evolve.

 Radar-based vehicle detection is one of the technologies that has continued to grow over the years. One of the advantages of the technology is the ability to track the speed and position of vehicles that navigate through an intersection approach. Speed and position values (i.e., vehicle trajectory data) are simply used to detect the presence of vehicles within a detection zone that emulates a loop and then discarded. Researchers recognized the value that the underlying discarded dataset used by radar-based vehicle detection system provides. For example, if properly logged and archived the dataset can be used to go back in time and understand detailed traffic conditions at a point in time and on a vehicle-by-vehicle basis. Detailed volume breakdowns by movement and lane, along with speed information, could then be used to objectively quantify the safety of individual intersection approaches. Researchers developed a data collection system to continuously monitor the underlying dataset of a commercially-available radar-based vehicle detection system without interfering with the operation of the signal controller. Algorithms to remove unwanted noise from the vehicle trajectories dataset were then developed and validated. The algorithms generate detailed volume and speed information, and other performance measures, that rely on actual field data to understand the traffic conditions at signalized intersections that use the radar-based vehicle detection system as a loop detector alternative.

Since thousands of intersection approaches across the country already rely on the radar-based vehicle detection system, the use of the data collection tools and algorithms developed open the doors to obtain unique insights into intersection safety by relying on actual vehicle trajectories and not on theoretical estimates or simulated scenarios.

Kelvin Santiago - is an Assistant Researcher at the Traffic Operations and Safety Laboratory from the University of Wisconsin-Madison. Kelvin received his Ph.D. in Civil and Environmental Engineering from the University of Wisconsin-Madison and his Bachelor of Science in Civil Engineering from the University of Puerto Rico-Mayagüez. His research ranges from the use of technology to tackle practical transportation engineering problems to the use of driving simulators to help designers conduct virtual road safety audits before the construction of a project. Kelvin is also an Adjunct Instructor at Madison College where he teaches Python computer programming courses focused on data analysis.



Session Material
Quick Links


Chris Osbourn


fb_logo   linkedin