Thesis Description & Objectives
Extended Reality (XR) technologies have become powerful platforms for immersive training, simulation, and entertainment. However, the quality and naturalness of user interaction remain constrained by tracking inaccuracies, hardware limitations, and individual differences in perception and comfort. These factors can lead to reduced precision, increased effort, or discomfort phenomena such as cyber-sickness, ultimately diminishing immersion and presence.
Recent advances in artificial intelligence and machine learning enable predictive approaches that analyze multimodal data to understand and anticipate user behavior. By processing inputs such as eye gaze, hand motion, physiological signals, application context, and user actions, AI models can infer interaction intent and detect variations in cognitive or physical state in real time. This predictive understanding allows the system to refine interactions dynamically, compensate for inaccuracies, and adapt both the behavior and appearance of interactive elements according to the user’s condition and task, resulting in a more stable and personalized experience.
This thesis aims to explore interaction refinement in XR through AI-based predictions of user state, interaction intent, and contextual conditions. Depending on the research direction, the focus may involve predictive enhancement of fine manipulation and control, or adaptive adjustment of interactive representations to mitigate discomfort and sustain immersion. The system will combine multimodal data from eye and hand tracking, user actions, and, when available, physiological sensors within an experimental XR testbed designed to evaluate adaptive interaction performance.
The research will include the design, implementation, and evaluation of an AI-driven framework for real-time interaction refinement. User studies will assess its impact on accuracy, comfort, and perceived presence. By leveraging AI to anticipate user intent and dynamically adjust both interaction and representation, this work aims to advance the development of intelligent, user-aware XR systems that provide natural, effective, and comfortable experiences across diverse scenarios.
References:
D. Calandra and F. Lamberti, “A Testbed for Studying Cybersickness and its Mitigation in Immersive Virtual Reality,” IEEE Transactions on Visualization and Computer Graphics, vol. 30, no. 12, pp. 7788–7805, Dec. 2024.
R. Islam, Y. Lee, M. Jaloli, I. Muhammad, D. Zhu and J. Quarles, “Automatic Detection of Cybersickness from Physiological Signal in a Virtual Roller Coaster Simulation,” IEEE VRW, 2020, pp. 648–649.
M. P. Olsson et al., “Predicting User Intent in Virtual Reality for Fine Manipulation Tasks,” Proceedings of the ACM Symposium on Spatial User Interaction, 2023.
D. A. Bowman et al., “Evaluation of Techniques for Predictive Interaction in Virtual Environments,” Proceedings of CHI, 2016.
