Signal Systems Testing

Background: A video detection system (VDS) is an application based on video processing technology that detects vehicles within a camera's field of view. A "tripline" VDS detects relative changes in the properties of the pixels located on the detection zones to identify the presence of a vehicle. VDS are increasingly being used, instead of inductive loops, at signalized intersections. The benefits of VDS include flexibility, adaptability to changing conditions, and their non-intrusive nature. In addition, VDS can be used for special purpose applications such as intersections at railroad crossings or detection on bridge decks. Additionally, some jurisdictions plan to use them to obtain traffic counts for redesigning timing plans and for signal coordination. Although VDS entered the vehicle detection market more than a decade ago, no previous studies have evaluated the systems from different manufacturers under specific weather and illumination conditions using side-by-side installation.

Research: This research (ICT project R43) compared the detection performance of three commercial VDS (Autoscope, Peek, and Iteris) at a signalized intersection under a wide range of illumination and adverse weather conditions. The results showed significant changes in the performance of the three VDS as illumination conditions changed from cloudy to sunny and from daytime to nighttime. Weather also affected the three VDS, generating different types of detection errors that could significantly affect the performance of actuated or adaptive traffic signals systems.

Impact: This study provided IDOT with the strengths and weaknesses of commercially available VDS compared to inductive loops. IDOT will use the recommendations to develop performance-based specifications to procure VDS, which will result in the purchase of functionally compliant VDS at competitive prices. Additionally, IDOT will use the insight gained to mitigate detection errors during adverse weather and thus maintain safe and efficient operations in these conditions.

PI: Ray Benekohal; Research Team: Ali Hajbabaie, Juan Medina, and Madhav Chitturi