Using Virtual Reality Techniques to Investigating Interactions Between Fully Autonomous Vehicles and Vulnerable Road Users
Samuel Labi, Professor of Civil Engineering – Purdue University
Sikai Chen, Bilsland Research Fellow of Machine Learning and Artificial Intelligence – Purdue University
Fully autonomous vehicles (FAV) or Level 5 automation vehicles can perform driving tasks in any environment and under all conditions without input from human drivers. However, they can lead to other challenges during real world implementation. Specifically, communication methods between vulnerable road users (pedestrians, bicyclists) and FAVs may change ultimately, which may lead to misunderstanding of intentions and cause more collisions. When road-users enter the road network, they initiate a constant exchange of information with the traffic environment and other road-users around them in order to be ready to respond immediately. Road-crossing pedestrians and bicyclists generally rely on informal communication methods, eye contact, facial expression and gestures, to interpret intentions of other road users and make decisions based on the information. With FAVs, these informal communication approaches cannot be realized. Hence, it is necessary to understand the interactions between these road users and FAV and design proper external features of FAV to establish efficient communication method. This project intends to expose participants to simulated testing environments with virtual reality headset and identify potential interface designs for FAV-pedestrian interaction. With pre-experiment collection of participants’ socio-demographic data, and behavioral measurements during the experiment, pedestrians’ attributes and factors that influence their road-crossing behavior and trust of AVs will be investigated.
Project Information Form