Trust in Autonomous Vehicles and AR Visualization

Abstract

This PhD project aims to enhance trust in autonomous vehicles by improving driver situational awareness through augmented reality (AR). Recognizing that unpredictable AI behavior and sudden transitions from autonomous to manual control can cause discomfort and safety concerns, the research will develop AR techniques to visually communicate the vehicle’s decisions and potential hazards to the driver. This approach will help drivers understand and trust the vehicle’s actions, especially in critical moments when manual intervention is needed. The project will tackle challenges in real-time hazard visualization and decision transparency, ultimately aiding the deployment of level 4 and 5 autonomous cars.

Thesis and defense

Thi Thanh Hoa Tran’s PhD (2022–2026) was a joint doctorate between IMT Atlantique and the University of South Australia. Its subject was Improving Autonomous Vehicle Drivers’ Confidence through Augmented Reality Visualization of Artificial Intelligence Decision-Making Context.

The thesis was defended on 16 February 2026, from 9:00 to 11:00, at IMT Atlantique’s Brest campus (B1-008). The official defense announcement includes the thesis abstract and the full defense information. The official thesis record and manuscript are available on theses.fr.

Research contributions

The thesis is structured around three complementary ways of designing AR information for autonomous-vehicle passengers.

1. Adding, removing, and modifying information

This contribution examines which driving-related and non-driving-related information should be displayed, altered, or removed to support trust in the vehicle.

Conference poster — IEEE VR 2025

Impact of Adding, Removing and Modifying Driving and Non-Driving Related Information on Trust in Autonomous Vehicles

View the IEEE VR 2025 poster entry · Read the poster paper

Full paper — unpublished manuscript (preprint / submission)

RAM-DRIVE: Remove, Add, Modify – Driving and non-driving Related Information for Autonomous Vehicle Trust Enhancement

2. Virtual character actions and emotional expressions

This contribution investigates how a virtual character’s actions and emotional expressions influence trust in an autonomous-driving scenario.

3. Highlighting, depicting, and predicting hazards

This contribution evaluates contextual visualizations that highlight relevant road users and depict predicted trajectories or time-to-contact information to improve trust and safety.

Published conference paper — ICAT-EGVE 2025

Full paper — unpublished manuscript (preprint / submission)

Combining Highlight and Depict Visualizations to Enhance Trust and Safety in Autonomous Vehicles

Related contribution: AR weather customization

During Taha Lamine’s CROSSING internship (February–August 2024), an additional contribution investigated whether passengers could customize AR-simulated weather conditions—rain, snow, sun, and clouds—and how this agency affects trust, acceptance, and experience in autonomous vehicles.

People involved

Researchers at IMT Atlantique

Guillaume Moreau

Professor of Computer Science at IMT Atlantique, Brest campus, in France, with research interests in Virtual and Augmented Reality within the Lab-STICC lab. He is also a member of the INRIA Hybrid team and currently serve as Deputy Dean for Research and Innovation at IMT Atlantique.

Researchers at University of South Australia

James Walsh

A believer in the technology’s ability to empower people and improve their lives, James is currently an Enterprise Fellow at the University of South Australia’s Australian Research Centre for Interactive and Virtual Environments (IVE). Having previously worked with education, Defence, the Australian Federal Police, as well as private industry across research, teaching, and commericalisation perspectives, James well versed in engaging with end users across all domains to deliver outcomes.

PhD Student

Thi Thanh Hoa TRAN

Thi Thanh Hoa Tran completed a Franco-Australian PhD within CNRS IRL CROSSING and the Lab-STICC INUIT team. Her research investigated how Augmented Reality visualizations of AI decision-making can improve situational awareness and trust in autonomous vehicles.

CROSSING intern

Taha Lamine

Taha Lamine was a CROSSING intern on this project and is now a PhD student at ENIB and Adelaide University within CNRS IRL CROSSING.

Research environment

This Franco-Australian doctoral collaboration was conducted within the CNRS International Research Laboratory CROSSING, which brings together research on interaction and Human / Autonomous Agents Teaming.