Summary of Projects Funded By The Federal Highway Administration for The National Surface Transportation Safety Center for Excellence From July 2006 to June 2014

An effort was undertaken to develop a system of image capture to analyze luminance data gathered in naturalistic driving research. Currently, still photometers exist to capture luminance data in an image as well as handheld luminance meters to capture luminance data in the environment. However, still photometers are unable to capture rapidly changing luminance data, while handheld luminance meters are time consuming to operate for multiple areas of interest within a scene and lack the ability to capture the scene's image.

The primary method used in this research was the calibration of smaller, more mobile luminance cameras through the use of an imaging photometer. Prediction results were compared to actual luminance recorded values, and an additional evaluation of data collection R&R was conducted. R&R evaluation results indicate the luminance camera is highly repeatable and capable of reproducing consistent data through multiple trials.

The calibration of a luminance camera for recording luminance data was found to be successful with a relatively high level of accuracy based on comparisons to known luminance values.

In nighttime driving, the visibility of a pedestrian or any object depends on its physical characteristics such as the luminance of the object and the contrast. The luminance, or light being reflected back to the observer, is a metric with little room for interpretation as to how it is derived. However, the concept of the contrast of an object has not been quite as clearly understood. The concept of contrast has been the focus of much investigation in past and current research.

This study compared multiple contrast metrics and implemented those determined to be most applicable as realistic measures. Specifically, researchers focused on the root sum of squares and power spectrum signature methods of determining contrast. These metrics were also used in conjunction with well-established metrics such as the Doyle and Michelson contrast metrics in order to make comparisons and inferences. These contrast metrics were also used in making comparisons to previously collected data of detection distances.

This project was an effort to develop a color camera system that could collect naturalistic video data with accurate color rendering.

Photometric devices can accurately measure color but cannot record the video data necessary for understanding visibility in dynamic environments such as nighttime driving. Video recorders can take video data yet are inaccurate in measuring color. To measure color and its effects on visibility in naturalistic settings, a color camera system was developed that could record video data with color rendering similar to what humans perceive. This system included a calibrated color camera and image analysis software.

The camera system was selected and calibrated in different lighting scenarios using a standard color chart. Calibration files were compared for color-rendering accuracy, and the best file, based on calibration in daylight, was selected for further analysis.

A new image-analysis method was developed. Along with its accompanying custom MATLAB programs, researchers were able to select portions of an image and analyze their three-dimensional color space coordinates. That capability would be useful in future work comparing photometric equipment and in analyzing naturalistic video data.

AVCs account for more than 5 percent of all reported motor vehicle collisions, and the problem continues to grow as both vehicle-miles traveled and wildlife numbers increase. Approximately 4 to 10 percent of AVCs involving large animals result in human injury. Costs for related property damage, medical care, crash management, and animal carcass management exceed $8 billion per year.

To mitigate this problem, an innovative roadside buried cable animal detection system was developed for testing at a particular site on the Smart Road where large wild animals are often observed. Researchers used surveillance systems to monitor animal movement and gauge system detections. Collected data were analyzed to determine the overall performance of the system and its suitability for implementation in problem areas on Virginia's public roads. This project was a partnership with VDOT.

Roadway images can detect and quantify adverse weather. Traffic cameras prevalent on highways and some secondary roads offer an opportunity to develop a system that detects adverse weather conditions automatically from the camera feeds and warn drivers using existing systems.

During the first phase of this project, the specifications for VDOT's latest digital cameras were identified; camera cost was determined with a consideration towards alternative, lower-cost cameras. Researchers also used an existing camera on the back of a VTTI building, similar to the current VDOT traffic cameras, capturing some photographs of the Smart Road at maximum zoom. These photographs were used to become familiar with the IFSTTAR modules and to create a Lambertian map of the scene. The IFSTTAR modules required specific image format and folder structures, so the team developed a method to convert the photographs and folders captured by the existing camera into data usable with the modules.

During the next phase, a brand-name camera identical to VDOT's current model will be installed on the Smart Road to collect data and develop prototype software using the IFSTTAR modules to measure fog visibility and rainfall rate. Also, a method of normalizing the data will be developed to eliminate the need for additional sensors for in situ calibration.

Most of the existing research on intersection lighting indicates the number of night crashes and the ratio of night-to-day crashes are lower at lighted intersections when compared with unlighted intersections. This research also suggests a safety benefit in the presence of lighting at rural intersections. Yet in all of these analyses, the presence of lighting is used as a binary variable: lighting is either present or absent. Consequently, the studies did not account for the lighting level, including the amount of light incident on the roadway surface (illuminance) or reflected from the roadway surface (luminance) from the roadway lighting system, or light from surrounding business establishments (e.g., parking lots, gas stations, malls, etc.). In addition, stray light from structures and establishments greatly affects the lighting level at both unlighted and lighted intersections. To better understand the importance of lighting at rural intersections, the relationship between lighting levels and night crashes at these locations needed to be studied.

Researchers conducted a comparative statistical analysis on lighting and crash data in order to understand the effect of lighting level and quality on the number of night crashes at rural intersections. This study analyzed rural intersections in the state of Virginia for the lighting level and crash information using the RLMMS. Adding this data to the crash data enhanced the quality and understanding of rural intersection lighting.

This project considered eye-glance and lighting issues in roundabouts because many State transportation departments are considering the implementation of roundabouts into their roadways. Pedestrian visibility at crosswalks, particularly on the roundabout exit, remains an issue. Investigating driver behavior and eye glances in the roundabout could provide insight into the optimization of lighting and crosswalk placement.

The study examined the performance of a driver in an active roundabout environment with double and single lanes. Participants were asked to drive through the roundabout with an eye-tracking device and a variety of signaling types. Eye-glance behavior was measured and linked to the lighting and signaling characteristics.

The results showed a variation in driver behavior at several aspects of the roundabout and are currently being used in the FHWA sponsored project Accelerating the Implementation of Roundabouts.