Skip to main content

University of Michigan School of Information


Stop, look and look out: Pedestrians' trust of autonomous vehicles at crosswalks

Pedestrian signal

Wednesday, 04/11/2018

The results of a new study show that pedestrians have higher levels of trust in autonomous vehicle (AV) driving behavior at signalized crosswalks, versus those that lack signals. 

Most studies measuring trust in AVs have focused on drivers and passengers, according to Lionel Robert, associate professor at the University of Michigan School of Information and principal investigator of the study. “Trust in AV: An Uncertainty Reduction Model (URT) of AV-Pedestrian Interactions” focuses on the behavior of pedestrians interacting with autonomous vehicles at crosswalks.

The study took place in an immersive virtual reality environment at the University of Michigan and was funded by Toyota Research Institute in Ann Arbor.  

The study manipulated the vehicle’s driving behavior to be defensive, normal or aggressive.  The traffic situation was crosswalks with and without signals. Researchers found that at crosswalks with signals, the driving behavior of the car had little impact on the pedestrian. But at crosswalks without signals, the AV’s driving behavior was a major determinant of trust. 

The study won third place in the “Best Late Breaking Report Award” category at the 2018 HRI Conference, which took place in March in Chicago. 

Co-principal investigators in the study are Anuj Pradhan, Xi Jessie Yang and Dawn Tilbury. Additional authors are Suresh Jayaraman, Chandler Creech, and Katherine Tsui.

Link to study

- Sheryl James, UMSI PR Specialist

Photo credit: Wikimedia Commons, downtowngal