Conversational and embodied AI agents in XR
Avatars and personas that listen, speak and act in mixed reality and in the browser, powered by speech and language models with full-duplex voice.
We combine extended reality with artificial intelligence and computer vision to build interactive systems that perceive, understand and respond to people in real time, and we take them out of the lab into healthcare, education, inclusion and industry.
AppiTech (Applied Interactive Technologies) is a research group of the Institute of Informatics at HES-SO Valais-Wallis, led by Prof. Antoine Widmer. We design immersive and intelligent systems (mixed and virtual reality, conversational avatars, body and gaze sensing) and evaluate them with the people who will use them: therapists, clinicians, teachers, students and patients.
We work in close collaboration with clinical, academic and industrial partners in Switzerland and Europe. Our projects run from co-design to field studies, and we build the open tools we need to measure what happens inside an immersive experience.
People stay in control: systems are co-designed with practitioners and evaluated in real settings.
From prototype to deployment on standalone headsets, in the browser, in the clinic and on the shop floor.
Self-hosted, open-source infrastructure where possible, so sensitive data stays under control.
Our work sits where extended reality meets AI agents that can perceive, reason, converse and act, while keeping people in control. It is organised around five interconnected areas.
Avatars and personas that listen, speak and act in mixed reality and in the browser, powered by speech and language models with full-duplex voice.
Understanding people, tasks and spaces in real time: body-worn cameras that follow a task as it is performed, gaze, room scanning and event-based sensing.
Mixed-reality and social-VR environments for therapy and prevention, designed with clinicians for children, adults and seniors.
Capturing what happens during an XR session (events, positions, gaze, conversations) to evaluate experiences and adapt them, with open tooling.
Bringing these systems to healthcare, health education, inclusion and industry, with attention to privacy, data sovereignty, accessibility and multilingual use.
A selection of the projects we lead or contribute to.
Immersive prevention centres that combine supervised group sessions in social VR with guided exercises at home, for two groups: young people with neurodevelopmental disorders and seniors with depressive or cognitive frailty. With Nantes Université (LS2N), CHU Nantes, the Institute of Psychology of the Polish Academy of Sciences, C2Care and DiverSSity.
A mixed-reality therapeutic platform for children with autism spectrum disorder or ADHD on Meta Quest: a magic garden, an aquarium, a fruit-slicing game, a party and a Snoezelen calming room, with a voiced gardener character who talks with the child.
Extended-reality simulations of work situations, such as a coffee break or a job interview, co-designed for young people on the autism spectrum. They talk by voice with expressive 3D characters and get feedback at the end.
An AI-based system that guides industrial workers step by step, in real time, through smart body cameras. It detects errors automatically, reduces downtime, and helps carry out complex tasks safely and efficiently.
Health students practise difficult conversations by voice with an AI-played patient or relative. Teachers write the scenarios without code; the platform follows the learning objectives live and gives feedback. Scenarios run in French, German and English, with speech in Swiss German, and in group scenes.
Greta, the gardener from DiverSSity, as a voice-driven avatar that converses in passthrough mixed reality and in the browser. It is our test bed for speech, expressive read-aloud and avatar actions, and it can present the lab itself.
An open-source platform for recording and analysing XR experiments. We run a custom version that adds a live headset view, room scans with body trails, and progress views for professionals. It logs the sessions of our own projects.
Professor of Computer Science · Head of AppiTech
His research combines immersive technologies with AI and computer vision to build interactive systems that understand and respond to human behaviour in real time, for healthcare, training and industry.
PhD student · MR & computer vision engineer
Works on immersive data for mental-health prevention and on mixed-reality development.
Mixed-reality developer
Develops XR4Inclusion and the lab's other mixed-reality projects.
Computer vision engineer
Works on industrial projects that use computer vision to help workers perform their tasks.
We welcome collaborations with clinical, academic and industrial partners, as well as Bachelor, Master and PhD students interested in XR and AI.