Project PI: Yunyi Jia
Project Duration: 03/2018-03/2022
Sponsor: National Science Foundation
Project Description:
This project addresses perceived safety within the context of autonomous driving. Self-driving cars have great potential to transform our transportation system, but their technical safety does not necessarily indicate how a human will perceive the safety of autonomous driving, especially if the driving style of the passenger is less aggressive than that of the self-driving car. The goal of this research is to change the car's controller when a human shows discomfort with respect to safety. The approach is to measure physical data from passengers in simulators and in real cars, compare these to expected levels of comfort, and update the controllers using novel algorithms. The project's results will improve user acceptance of autonomous vehicles, as well as produce new curriculum and outreach programs related to autonomous driving.
The goal of the project is to understand human perceived safety and comfort in autonomous driving and examine and bilaterally adapt the interactions between humans and autonomous vehicles in order to enhance the human perception (i.e., perceived safety and comfort) while simultaneously enhancing the vehicle performance (i.e., technical driving safety and comfort). In particular, the proposed research will (a) understand the factors affecting the human perceived safety and comfort and investigate how to quantify human perceived safety and comfort using physiological signals, (b) investigate general human and autonomous driving models and online bilateral adaptation strategies to improve the driving safety and comfort, and (c) evaluate the research outcomes using both simulated and real autonomous vehicles. The research will contribute to improving the safety and comfort of autonomous driving and is scalable to other human-autonomy interaction contexts as well.
Publications:
H. Su and Y. Jia*, "Computational Modeling of Human Comfort in Automated Vehicles from Maneuvering Behaviors," IEEE Intelligent Transportation Systems Conference (ITSC), 2022.
J. Xiang, L. Guo and Y. Jia*, "Comfort Improvement for Autonomous Vehicles using Reinforcement Learning with In-situ Human Feedback," SAE Technical Paper, 2022.
H. Su and Y. Jia*, "Study of Human Comfort on Autonomous Vehicles Using Wearable Sensors," IEEE Transactions on Intelligent Transportation Systems, pp. 1-15, 2021.
L. Guo and Y. Jia*, "Anticipative and Predictive Control of Automated Vehicles in Communication-Constrained Connected Mixed Traffic," IEEE Transactions on Intelligent Transportation Systems, pp. 1-14, 2021.
L. Guo and Y. Jia*, "Inverse Model Predictive Control (IMPC) based Modeling and Prediction of Human-Driven Vehicles in Mixed Traffic," IEEE Transactions on Intelligent Vehicles, pp. 1-12, 2020.
L. Guo and Y. Jia*, "Predictive Control of Connected Mixed Traffic under Random Communication Constraints," IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2020.
Y. Jia* and B. Ayalew, "Cyber-Human-Physical Heterogeneous Traffic Systems for Enhanced Safety," IEEE International Conference on Connected and Autonomous Driving, 2020.
D. Bolduc, L. Guo and Y. Jia*, "Multi-Model Approach to Personalized Autonomous Adaptive Cruise Control," IEEE Transactions on Intelligent Vehicles, vol. 4, no. 2, pp. 321-330, 2019.
L. Guo and Y. Jia*, "Modeling, Learning and Prediction of Longitudinal Behaviors of Human-Driven Vehicles by Incorporating Internal Human Decision-Making Process using Inverse Model Predictive Control," IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2019.
X. Wang, L. Guo and Y. Jia*, "Road Condition based Adaptive Model Predictive Control for Autonomous Vehicles," ASME Dynamic Systems and Control Conference (DSCC), pp. V003T37A005, 2018.
Dataset:
Clemson Comfort Dataset: http://cecas.clemson.edu/comfort/