Computer software being developed at Iowa State University may give traffic safety engineers clues to why drivers ignore traffic signals at intersections.

Derrick Parkhurst, an assistant professor of psychology and human computer interaction, and Nathan Willis, a sophomore from Omaha majoring in aerospace engineering, are working on a computer program to automatically analyze video taken at intersections. A camcorder mounted on a traffic sign or signal records vehicles in motion, including instances in which drivers fail to stop completely at a stop sign. Computer software analyzes the footage to measure the frequency of unsafe driving and determine potential causes.

The Iowa Department of Transportation is supporting the ISU project with a $10,000 grant in hopes that the analysis will help them devise ways to reduce the occurrence of traffic accidents.

Jamie Reyes, a safety review engineer with the Iowa Department of Transportation, likes the concept, and believes that it’s time to take a “behavioral approach” to analyzing driver behavior.

“Bigger stop signs and flashing lights are not the answer,” Reyes said. “I believe that we have exhausted the engineering aspect of solving highway problems, and that we are now in the realm of driver behavior. Even if we put a lot of bells and whistles out there at intersections, but we still have accidents.”

To design a safer highway system, Reyes said, he thinks it would help to analyze cases in which drivers ignore the law to see if any contributing factors can be identified, such as whether they are talking on their cellular phones, if they appear to be distracted and how fast they’re moving through the intersection.

Parkhurst and Willis consulted Reyes before beginning their research on driver behavior last spring. Willis took on the project through the freshman honors program, with Parkhurst as his mentor. The pair originally looked at projects relating to Willis’ major in aerospace engineering, but instead decided to explore automobile transportation issues to address the IDOT’s problem.

Willis earned college credit by working on the project during the spring semester last year, and received an honors research grant to work on it full time over the summer. He said he still spends about 10 to 15 hours per week working on it now. Parkhurst estimates that the project is about halfway complete. Although the software appears to work, more testing will be needed to finalize it.

“With this route, we could possibly help the DOT, and it was a way to incorporate human computer interaction, which I was interested in learning more about,” Willis said. “It’s been an eye-opening experience that has made me see the potential for human computer interaction in aerospace or any other discipline.”

Human computer interaction is the study of the relationship between people and computers. ISU introduced a graduate program in this field in 2002, and is investing resources to become a leader in the discipline.

“Computers are at the point where they can help predict and monitor your behavior,” Parkhurst said. “Right now, traffic safety data is based off accident data, when we can get a more accurate picture of what’s going on at an intersection by monitoring it for a month with a low-cost camcorder and using our software to detect specifics about the drivers. That’s the development we’re hoping for down the road.”