Skip to main content

Researchers study how humans respond to autonomous driving cues


Research is being conducted at the Center for Automotive Research (CAR) on how human drivers respond to a driving speed suggestion.

Driving in an autonomous vehicle simulation

The data is being collected in a driving simulator at CAR’s Driving Dynamics Lab (DDL). Assistant Professor Stephanie Stockar is leading this project along with Research Scientist Jeff Chrstos who leads the DDL and Professor Marcello Canova. Graduate students James Farrell and Luis Mendez, visiting scholars Marcello Telloni and Radu Pirvan, Research Associate Mehmet Ozkan and DDL graduate student Gourav Vaid are also involved in the project.  

Although autonomous vehicles are on the rise, it will still likely be a few years before they deploy on a large-scale. In the meantime, Stockar and her team are working on a semi-autonomous vehicle that has some features of autonomy, yet still relies on human intervention. These vehicles may suggest or implement certain speeds autonomously, and the team is trying to figure out if suggesting a speed profile helps the person or does more harm than good.

These speed suggestions may bother a human driver if they don’t know why the vehicle is braking or accelerating. Some people may also not agree with the vehicle’s speed suggestions; some may be comfortable with more aggressive speed suggestions, whereas others prefer a smoother, slower approach to driving. So, how do human drivers respond to a driving speed suggestion?

Stockar and her team are bringing in human subjects to answer this question. The team will ping suggestions to the subjects while they are in the driving simulator to see how they respond. Studying a wide variety of drivers will give them a more holistic understanding of how people respond to speed suggestions.

Stockar was inspired by her work on the ARPA-E NEXTCAR project, and this project in the DDL aims to bridge the gap in technology between the NEXTCAR project and the current status of autonomous vehicle implementation. Stockar says, “We’re trying to understand what the ideal way is of suggesting a speed profile that a driver is comfortable following. We know from NEXTCAR that if we can get the driver to follow the suggestions, we can get major improvements in fuel economy and vehicle electric range. This, however, requires building trust between the human operator and the vehicle.”

Written by CAR Writing Intern, Cassie Forsha

Category: Research