How to Share the Road Safely with Automated Vehicles

If you’re a pedestrian at a crosswalk, how do you know if an approaching car is really going to stop? As a driver at a 4-way stop, can you determine who yields the right-of-way? In these situations, people use visual cues, auditory cues, and a degree of prediction based on traffic laws and social norms to determine the safety of their next moves. But how do you figure out the “intentions” of an automated vehicle (AV)?

Westat is seeking the answer to this question in a project for the National Highway Traffic Safety Administration (NHTSA). The 2-year project will help identify ways that highly automated vehicles can be designed so they interact well with other road users.

“We’re assessing the human aspects of interacting with automated vehicles,” says James Jenness, Ph.D., a Westat Associate Director and project director for the study. “Specifically, we’re determining what information AVs must communicate to shared road users to eliminate uncertainty and avoid accidents.”

Identifying Cues to Predict Behavior

Westat conducted an on-the-road study to identify the cues people need and use to predict behavior. Study participants were recruited and trained to provide think-aloud commentary while traveling through various traffic scenarios where crashes are common. These include intersections, lane merges, and parking lots.

Pedestrians in this study were outfitted with synchronized audio- and video-recording cameras while the same equipment was mounted on bicycles and cars. Each participant recorded their perceptions of the intent of other vehicles and drivers, and when and how they decided to proceed through traffic.

Measuring Reaction to Signals

Westat also conducted a human factors lab study with 30 participants to determine if people can understand and react appropriately to different AV signaling systems. Participants watched a life-size video of an automated vehicle coming at them. The video was manipulated with several different overlays to indicate possible signaling system concepts, such as lights on the top or front of the car. Participants answered a series of questions to indicate how well they understood each of the signals.

Finding Answers to Enhance Safety

Westat’s findings will help NHTSA identify and recommend the most effective ways AVs can communicate intent. The research will also support government agencies’ creation of safety programs and messaging related to automated vehicles.