Intelligent Avatars for Collaborative and Rehabilitative Extended Realities

About the Project

This project explores the development of intelligent, embodied avatars in Extended Reality (XR) environments, aimed at supporting collaborative and rehabilitative tasks. By integrating conversational AI with generative models for human motion, these avatars will respond naturally to users through voice, gesture, and movement. The avatars must be interactive and capable of reacting to the human user’s actions in real time, enabling immersive, adaptive interactions that enhance engagement, motivation, and task performance.

Current XR systems often rely on pre-scripted or rigid avatar animations, limiting realism and responsiveness. Recent advances in generative AI – especially deep learning models trained on large-scale motion capture data – now make it possible to synthesise realistic, diverse, and context-sensitive human motion in real time. This capability is essential for creating avatars that move in ways consistent with their speech, context, and role – whether demonstrating a rehabilitation exercise or guiding users through a collaborative task. Lifelike and situation-aware motion is particularly impactful in therapeutic and learning environments, where user perception, emotional engagement, and comprehension are closely tied to avatar believability.

Candidates should have appropriate academic qualifications (first or upper second class honours, and preferably MSc) in Computer Science, Engineering, Mathematics, Physics or other relevant area, strong background in programming and desire to become experts in Artificial Intelligence and Extended Realities.

Funding Notes

there is no funding for this project

References

• Hixon-Fisher O, Francik J, Makris D., 2025, February. Pose-centric motion synthesis through adaptive instance normalization. Proc. of 20th Int. Joint Conf. on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISAPP), Porto, Portugal (pp. 39:47).
• Nagy, B., Prieto, E.H., Mirza, A., Johnson, G., Usman, M.A., Smith, C., Francik, J., Bakirtzis, C., Politis, C. and Grigoriadis, N., 2025, January. Therapy reloaded: mixed reality games for hand rehabilitation. In 2025 IEEE 22nd Consumer Communications & Networking Conference (CCNC) (pp. 1-6). IEEE.