Thesis Proposals

Reset

Collaborative multi-LLM agents in eXtended Reality for human-centered task support

Advisors: Alberto Cannavò, Fabrizio Lamberti
Problem: Recent research highlights that XR + LLM integration is still fragmented, lacking unified frameworks for interaction and collaboration. Human-AI collaboration benefits from complementary reasoning between humans and AI, but current systems poorly coordinate roles and interaction dynamics. Moreover, existing XR systems typically treat AI as a tool, not as a team of collaborating agents, and do not support multi-agent coordination alongside humans in shared virtual environments. Objective: Design, implement, and evaluate a human-centered XR framework where multiple specialized LLM agents collaboratively assist users in complex tasks within a shared virtual environment. Specific goals include: i) Enable multi-LLM collaboration; ii) Integrate human-in-the-loop interaction within XR iii) Support real-time co-presence and co-decision-making; iv) Evaluate improvements in task performance, cognitive load, and user trust. Approach: The proposed system will include: i) an immersive workspace based on XR for interactions; ii) multi-LLM agent system and iii) Human user as supervisor and collaborator. Possible use cases may consider, e.g., collaborative design; medical or scientific decision-making, training and education in XR labs. Expected Impact: Improving user performance and experience.
Target Programs: Ingegneria del Cinema e dei Media Digitali | Ingegneria del Cinema e dei Mezzi di Comunicazione | Data Science and Engineering | Ingegneria Informatica (Computer Engineering)
Keywords: AGENTIC AI, AUGMENTED REALITY, COLLABORATION, HUMAN-MACHINE INTERACTION, LLMS, VIRTUAL REALITY

Automated 3D Digital Twin generation from video sequences for eXtended reality (XR) simulation

Advisors: Fabrizio Lamberti, Gabriele Pratticò, Lorenzo Valente
Context and Motivation: In the era of Industry 4.0 and the Metaverse, Digital Twins (DT) have become indispensable for real-time monitoring, predictive maintenance, and immersive training. However, the creation of high-fidelity Digital Twins remains a significant bottleneck. Traditionally, this is a manual, labor-intensive process—especially when the object’s functional behavior must be modeled and coded from scratch to match its physical counterpart. Problem Statement: Current Digital Twin generation methodologies focus primarily on static geometry. However, most industrial assets and consumer products are Cyber-Physical Systems (CPS) or articulated objects (e.g., hardware tools, machinery with levers, knobs, and buttons) characterized by specific functional behaviors. For Extended Reality (XR) simulations to be effective, a Digital Twin must exhibit "Physical-Behavioral Symmetry": it must not only look like the physical object but also mimic its kinematic constraints and logic. Manually defining multi-part hierarchies, kinematic joints, and Finite State Machine (FSM) logic for every asset is time consuming and require specialized skills (3d modeling, programming). Objectives : The objective of this thesis is to propose an end-to-end pipeline that reconstructs articulated/cyber-physical 3D models from standard video sequences and automatically synthesizes the underlying code/logic required to drive their behavior in virtual environments (e.g. Unity application). Innovation and Expected Impact: Students may leverage existing technologies as a starting point—such as PartGen for part-based reconstruction and LLMs for code generation—to either extend these frameworks or propose a novel, integrated Machine Learning-based approach. The successful outcome of this thesis will be a framework capable of transforming a simple videos into a "Smart Asset": a ready-to-use, interactive Digital Twin for XR that possesses both geometric fidelity and functionality.
Target Programs: Ingegneria del Cinema e dei Media Digitali | Ingegneria del Cinema e dei Mezzi di Comunicazione | Data Science and Engineering | Ingegneria Informatica (Computer Engineering)
Keywords: 3D RECONSTRUCTION, CODE GENERATION, DIGITAL TWIN, FOUNDATION MODELS, ROBOTICS, VR, AR, 3D, CG, HCI
💼 Thesis In Collaboration With a Company

Augmented Reality support for aircraft wing de-icing operations

Advisors: Fabrizio Lamberti, Gabriele Pratticò
Aircraft wing de-icing is a critical safety procedure in winter operations, as even thin layers of ice or frost can severely compromise aerodynamic performance and flight safety. In real operational environments—characterized by low visibility, time pressure, and complex aircraft geometries—operators may experience difficulties in accurately identifying residual ice or areas that still require intervention. This thesis is part of a project funded by Regione Piemonte, in collaboration with companies specializing in drones and computer vision, aiming to develop an Augmented Reality (AR) system that enhances current de-icing workflows. The proposed solution will integrate drone-based imaging and vision algorithms developed by other partners to detect critical areas and visualize them in AR to support operator decision-making. The system will be collaboratively developed and experimentally evaluated together with SAGAT at Torino Caselle Airport. The thesis is addressed to students in Computer Engineering with a focus on graphics and interactive systems, as well as to students in Cinema Engineering and Aerospace Engineering. The thesis is meant to be developed at Politecnico di Torino with experiences on-field at the airport, though one of the involved companies could also be available to host the student.
Keywords: AIRCRAFTS, AIRPORT OPERATIONS, AUGMENTED REALITY, VISUALIZATION

Adaptive cybersickness mitigation in Virtual Reality based on real-time user state detection

Advisors: Fabrizio Lamberti, Davide Calandra, Gabriele Pratticò
Cybersickness is a disorder characterized by symptoms such as nausea or discomfort that can occur during or after the use of Virtual Reality (VR) technologies. Similarly to motion sickness, it is mainly caused by inconsistencies between visual and auditory stimuli from the simulated environment and the expected feedback from the vestibular system. Several mitigation and prevention techniques have been proposed and are employed in various commercial products. However, most of them are static, highly situational, or require design choices that negatively affect key aspects of the VR experience such as immersion, naturalness, and sense of presence. A promising direction that remains underexplored is the use of adaptive mitigation techniques that respond dynamically to the user's real-time state. Recent advances in physiological monitoring and data-driven modeling make it possible to estimate cybersickness symptoms during VR exposure by analyzing signals such as heart rate or skin conductance, as well as behavioral data like movement patterns and interaction metrics. These indicators, when interpreted using machine learning or rule-based models, offer the opportunity to adapt the virtual environment continuously in response to the user's current level of discomfort. The objective of this thesis is to design and implement an adaptive mitigation system capable of modifying VR parameters in real time based on continuous detection of cybersickness indicators. The system will be integrated into an existing VR environment and will be tested by adjusting elements such as motion intensity, visual effects, or field of view in order to alleviate symptoms without significantly compromising immersion. The effectiveness of the approach will be evaluated through a user study, combining subjective and objective measures to assess comfort, presence, and user experience over time. References: - Jyotirmay Nag Setu et al., Mazed and Confused: A Dataset of Cybersickness, Working Memory, Mental Load, Physical Load, and Attention During a Real Walking Task in VR, 2024 IEEE International Symposium on Mixed and Augmented Reality (ISMAR), 10.1109/ISMAR62088.2024.00121 - Davide Calandra, Fabrizio Lamberti, A Testbed for Studying Cybersickness and its Mitigation in Immersive Virtual Reality, IEEE Transactions on Visualization and Computer Graphics ( Volume: 30, Issue: 12, December 2024), 10.1109/TVCG.2024.3448203 - Rifatul Islam et al., Cybersickness Prediction from Integrated HMD’s Sensors: A Multimodal Deep Fusion Approach using Eye-tracking and Head-tracking Data, 2021 IEEE International Symposium on Mixed and Augmented Reality (ISMAR), 10.1109/ISMAR52148.2021.00017
Target Programs: Ingegneria del Cinema e dei Media Digitali | Ingegneria del Cinema e dei Mezzi di Comunicazione | Data Science and Engineering | Ingegneria Informatica (Computer Engineering)
Keywords: ARTIFICIAL INTELLIGENCE, HUMAN-MACHINE INTERACTION, MACHINE LEARNING, VIRTUAL REALITY
12