As automated driving systems begin to operate on pubic roads, there is a need to understand how pedestrians and other road users interact with vehicles based upon their perceptions of vehicle behavior, so that automated vehicles can be developed to be as predictable as possible. Interactions between automated driving systems and people constitute a network of potential users including bicyclists, passengers, pedestrians, other road users and remote operators.
How do vehicle behaviors influence the perceptions of riders’ or other road users’ interaction with driving automation?
What We Did
Automated driving systems need to consider and respond to the needs of all road users interacting with them in order to foster trust and acceptance in society. One way that existing vehicles communicate driver intentions is through on-road behaviors, such as braking and acceleration profiles. These vehicle behaviors can influence people’s perceptions of politeness, safety and ultimately trust in automated driving systems.
This project has two main focuses:
1) How bicyclists coordinate with full and partially automated vehicles.
2) How other road users coordinate with fully automated vehicles with social interactions.
The findings of this research project are expected to improve the overall interactions between automated driving systems and pedestrians and other road users. Preliminary results indicate the following:
- Stopping short of a crosswalk so that the vehicle indicates it is yielding to a pedestrian was found to be a frequent and reliable communication tool between drivers and pedestrians in naturalistic observations.
- Braking profiles — such as conservative, moderate and aggressive — affect a driver or passenger’s tendency to seek a method to influence the vehicle’s behavior (pressing a pedal or grasping the steering wheel).
This is a project in collaboration with the University of Wisconsin