Exploring the Potential of the Industrial Metaverse in Systems Engineering: Hope Beyond the Hype?

Sergio Camilo MEDINA GALVIS, Romain PINQUI, Frédéric NOËL

Collaborative Virtual Environments (CVEs) with adaptive mechanisms are essential for addressing the complexity of engineered systems design tasks involving a multidisciplinary team of experts modeling different views of the system with their preferred visual metaphors and devices. This research focuses on self-morphing CVEs with three key features: multi-view, multi-representation, and multi-device for visualization and interaction. This Ph.D. thesis aims to conceptualize, demonstrate, and evaluate a self-morphing CVE that supports synchronous and asynchronous collaborative systems architecture tasks. This position paper invites discussion on metrics for assessing adaptive CVEs and the impact of device-specific features on collaboration during system modeling activities.

The Role of Telepresence in Consumer Environmental Responsibility: The Case of Virtual Reality Educational Experiences

Caroline BONNETIER

MATCH: a Mixed-Reality Cognitive Assistance Technology to Support Independence at Home for Older Adults with Neurocognitive Disorders

Guillaume SPALLA, Charles GOUIN-VALLERAND, Nathalie BIER

We present MATCH, a mixed-reality assistive technology for cognition to support independence at home of older adults with neurocognitive disorders. We designed, developed and evaluated MATCH following a human-centered design approach. This paper presents an overview of the methodology used for this project, the results and a discussion.

The Use of Virtual Reality in Public Speaking Training: Design of a Dedicated Tool

Sarah SAUFNAY, Elodie ETIENNE, Michaël SCHYNS

Given their significance in many management contexts, public speaking skills are increasingly valued by organizations. Not always innate, these skills can fortunately be acquired through practice. However, due to their intrinsic social aspect, opportunities to refine these skills in real-life settings are scarce. Faced with such limitations, Virtual Reality emerges as a suitable and innovative solution. By simulating public speaking scenarios, it allows users to practice in front of an audience, albeit a virtual one. The added value of Virtual Reality public speaking tools is therefore clear, but the optimal training approach and environment design are not of a basic nature. In order to identify best practices, a dedicated Virtual Reality training system has been developed and will be presented in light of the scientific literature. The objective of the present research project is to leverage the system to investigate unanswered questions pertaining to the perception of virtual audiences, the evaluation of public speaking performance, and the effectiveness of such tools.

Understanding the Influence of Visual Metaphors on Cognitive and Transdisciplinary Collaborative Conceptual Co-Design Activities

Ghislain MUGISHA, Romain PINQUIE, Emilie LOUP-ESCANDE

This research, at the beginning of a doctoral thesis, focuses on the critical yet ambiguous conceptual design (CD) phase in systems engineering. While Model-Based Systems Engineering (MBSE) is increasingly used, its reliance on 2D symbolic notations like SysML hinders acceptance and collaboration among multidisciplinary stakeholders. This study explores how 2D/3D visualizations combined with symbolic/iconic and stereoscopic/monoscopic devices influence cognitive and collaborative activities during transdisciplinary co-design. By addressing challenges of engagement and comprehension across experts and non-experts, the research aims to provide insights into designing more inclusive and effective external representations in MBSE.

Optimizing Social Interactions in VR

Eloïse MINDER, Solène NEYRET, Sylvain FLEURY, Christophe GUILLET, Jean-Rémy CHARDONNET

Technological developments have made Virtual Reality (VR) technologies accessible, and have democratized its use for industrial, cultural or entertainment purposes. VR can be seen as a unique medium: while being immersed together in a virtual environment (VE), people can feel the presence of each other as if they were in the same physical space even if they are far apart in reality. This phenomenon is known as co-presence. The VR medium introduces specific factors influencing on the dynamics of the social interaction. This PhD project aims at understanding that dynamics through the study of co-presence.

Leveraging Generative Artificial Intelligence for Mixed Reality Educational Activities

Mohammed Oussama SEDDINI, Mohamed EZZAOUIA, Iza MARFISI

This article explores the potential of generative Artificial Intelligence (AI) to facilitate the creation of Mixed Reality (MR) educational activities. Through a prototype integrating three AI modules (AI Prompter, AI Chat, and AI Tutor), the system enhances the design process by generating multimodal content, proposing activities based on existing teaching material, and offering personalized and immersive experiences to learners. The prototype was evaluated during a preliminary study with a teacher. The results show promising interest in these tools, while highlighting technical limitations requiring future improvements.

Systematic Design and Evaluation of Immersive Environments to Support Project in Living Lab Mode

Anaëlle HILY, Laurent DUPONT, Mauricio CAMARGO, Jérôme DINET

The application domains of Extended Reality and more specifically virtual reality (VR) have expanded rapidly, driven by the advancement and accessibility of immersive technologies. This growth has led to an increase in their use in both academic research and industry, showing concrete benefits in training, raising awareness in emergency situations, and providing remote collaborative spaces. In research, VR environments offer experimental advantages as well, notably due to precise control over variables, which offers better reproducibility. However, VR is not limited to its technological components. It is a multidisciplinary field that requires expertise in cognition, computer science, and engineering. Developing new VR applications presents a complex challenge, where integrating user feedback is crucial to mitigate potential negative effects. Despite the field’s growth, few theoretical frameworks exist to guide the design process, and validation often relies on subjective self-reported questionnaires providing partial data on user experience (UX). This paper introduces part of our work to help formalize a design approach for VR systems, incorporating a comprehensive evaluation of UX through a combined methodology of subjective and objective indicators, including physiological measures. Conducted through a multidisciplinary collaboration, this research implements an action research approach to address the various levels and disciplines involved in VR system development.

AVATOUCH - Synchronous Cooperation Based on Social Intention for the Manipulation of a Virtual Object in a Collaborative Virtual Environment

Mattéo BORREGO

The project focuses on cooperation for the manipulation of a virtual object by two individuals in remote immersive environments, without verbal communication or instruction, relying solely on gestures. Within the framework of the PEPR eNSEMBLE program, our aim is to strengthen this cooperation by drawing on theories from cognitive sciences, particularly the concept of intentionality in a social context. Indeed, during social interactions, individuals communicate their intentions and emotions through a variety of nonverbal cues. Among these, facial expressions, gazes, and gestures significantly contribute to the transmission of information that is essential for the success of social interactions. Moreover, inspired by the literature on social touch between humans and recent research on human-agent interaction with haptic feedback, we have worked at Heudiasyc on the sensation of being touched by an agent (ANR Socialtouch, ANR Match) in collaboration with ISIR (Catherine Pelachaud) in a virtual environment. These studies were conducted using virtual reality headsets, but to our knowledge, none have yet explored social touch between remote immersive rooms, involving life-sized user avatars. Therefore, the objective of this thesis project is to explore cooperation between remote users, represented by their virtual bodies (avatars) with multisensory feedback and intentional gestures, by adapting the paradigm of Quesque et al. (2013), with and without social touch, through a set-up consisting of two connected immersive rooms that are part of the CONTINUUM project.

Toward EEG Discrimination of Fingers Movements during Motor Imagery vs Passive Movement

Théo LEFEUVRE, Kyungho WON, Monica MALVEZZI, Mihai DRAGUSAN, Claudio PACCHIEROTTI, Anatole LECUYER, Marc J-M MACE, Léa PILLETTE

This study investigates the potential of passive finger movements as an alternative to motor imagery (MI) for the calibration of brain-computer interfaces (BCIs). Twenty-six participants performed both MI tasks and passive movement tasks involving finger flexion and extension, while their electroencephalographic (EEG) brain activity was recorded and analyzed. Brain activity in the ω (8–12 Hz) and ε (20–25 Hz) bands was significantly stronger during passive movement tasks compared to MI tasks. Additionally, task conditions and movement types had independent and significant effects. These findings suggest that passive movements could improve the accessibility and reliability of BCIs, particularly in applications requiring precise finger control such as extended reality applications.

Robust and Efficient AI Motion Capture

Georgios ALBANIS, Nikolaos ZIOULIS, Spyridon THERMOS, Anargyros CHATZITOFIS, Kostas KOLOMVATSOS

This document provides an overview of our research on developing robust and efficient AI-based Motion Capture (MoCap). While recent AI MoCap solutions have shown impressive results (Goel et al., 2023), they often require extensive post-processing and cleanup to produce usable assets, enabling downstream applications. These challenges arise from factors such as noise in 2D keypoint estimators, mismatches between predictions, and jitter in the output. Our work addresses these issues by introducing robust and temporally-aware neural pose solvers that maintain computational efficiency. We demonstrate that this approach can yield ready-to-use assets with minimal additional processing.