Head
- Manager: Hedi TABIA (PR Uni. Evry)
- Assistant manager: Amine CHELLALI (MCF Univ. Evry)
Field of activity
The IRA2 team’s research focuses on the design, production, and evaluation of systems to assist people in their environment (indoors or outdoors). The environment’s diversity, complexity, and unpredictability require the use and control of complex artificial systems (sensors and heterogeneous robots) to provide such assistance. Research work is implemented and evaluated on various platforms, including EVR@ and ARPH platforms. The aim is to remove the technological and scientific barriers to perception, interaction, and decision support for people. The aim is to propose :
– objective 1: multisensory and collaborative assistance for interaction and decision-making for people immersed or semi-immersed in a virtual world (virtual reality).
– objective 2: personal assistance in direct interaction with the augmented real environment (augmented reality).
– objective 3: functionalities for a “robot assistant” cooperating with the operator and evolving in a home automation environment (ambient robotics).
The field of application of these assistances is highly diversified. It concerns both able-bodied and disabled people. The person can be assisted at home, in an institution, on the move (pedestrian or motorized mobility), or in a professional environment.
Mixed reality interaction and collaborative work
Mixed Reality Interaction (MRI): Research is focused on assisting 3D interaction in environments that can be real-augmented or virtual. In this line of research, particular attention is paid to how we can help people to interact (navigate, select, manipulate and control the application) in an effective and credible way with their environment. This thinking has led us to propose software models and techniques for assisting 3D interaction in mixed reality. Work in this area began with the acquisition of the EVR@ virtual and augmented reality platform in 2004, which is essential for the study and evaluation of 3D mixed reality interaction techniques.
The scientific challenges being addressed are: 1) the latency of human-system interaction; 2) the performance and credibility of 3D interaction techniques in relation to the task; 3) the adaptability of 3D interaction in relation to virtual reality interfaces and the context of the task (context-sensitive interaction). The latter is being addressed in collaboration with IBISC’s SIMOB team.
Collaborative work (CW): Research is motivated by the use of synchronous groupware and, in particular, the use of VCAs (Collaborative Virtual Environments) by several people simultaneously. In this line of research, particular attention is paid to the way in which a collaborative activity can be structured, organized, managed and adapted. This has enabled us to propose formalisms and models to support collaborative activity (communication, coordination and co-production).
The scientific challenges addressed are those linked to: 1) the malleability of collaboration systems (groupware) to take into account new needs that emerge during user collaboration; 2) the coordination of group interactions (collaborative interaction) in VCAs (Collaborative Virtual Environments).
All this work has given rise to several publications and research project deliverables: ANR RIAM DigitalOcean 2006-2009, European FP6 VENUS 2006-2009, European FP7 RSME DigitalOcean 2011-2012.
Augmented Reality (AR)
The research activity of this theme, which aims to satisfy the second objective stated in the general problematic, is to design systems capable of making a virtual world coexist spatially and temporally with the real environment. The aim of this coexistence is to enrich the user’s perception of his real environment through visual, sound or haptic augmentations. The environment can be either indoor or outdoor. The user may be present in the real environment, or may perceive it from a distance. If the user is present on the actual site, visual augmentations can be made either by direct vision (wearing see-through glasses) or by indirect vision on a screen.
We positioned ourselves early on in the field of AR. Our research focuses on the modeling and calibration of hybrid sensors, 3D prediction, and registration, and robust real-time hybrid object recognition and tracking in the presence of occlusions. One of the strengths is the development of robust hybrid tracking techniques (tracking based on visual and inertial information of objects in the presence of occultations).
AR is inherently multidisciplinary, which is reflected in the IT developments associated with this field of research. To develop an application in this context, we need to consider requirements specific to the field and generic requirements linked to all IT development.
These requirements are classically divided into functional requirements (linked to the application’s core business) and non-functional requirements (transversal and not linked to the core business). It’s the latter category that we’re focusing on to facilitate the work of the developer, who can then concentrate on the business side (for AR, these include image processing algorithms, localization, and registration issues, as well as the design of solutions enabling interaction in a Mixed Reality space).
Initially, we focused on providing modular, standardized solutions based on component programming to facilitate developments in a field where sensors and effectors are sometimes supported experimentally and where algorithms evolve rapidly. This programming paradigm has proved useful and facilitates the rapid prototyping of AR applications, while enabling components to be simply substituted for others when one of the associated techniques or technologies evolves.
The results obtained in this first phase were used in the ANR project RAXENV (Outdoor augmented reality applied to the environment ANR06-TLOG14 2007-2010).
Mixed reality interaction and collaborative work
The Ambient Robotics theme focuses on the contribution of robotic methods and technologies to the needs of the disabled and elderly. The robot assists the person and/or his or her entourage, either autonomously or by being remotely operated by a local or remote user.
Our positioning is original in relation to the assistive robotics community, as we believe that autonomous robotics is not sufficiently mature to respond to the societal problem posed by loss of autonomy. Faced with the diversity of situations, users, robots, and tasks, we propose a “variable geometry” response based on providing the user with control modes ranging from assisted teleoperation (in which the operator is assisted) to assisted autonomy (in which the robot is assisted). The need for a variety of control modes has been confirmed by the experimental results of the QuoVadis project (ANR TecSan, 2009-2012), the ANR RobautiStic project (2007-2010), and Companionable (European IP, 2008-1013). The answers we provide must respect the medical, ethical and socio-economic constraints specific to this field of application. This is why our projects are highly multidisciplinary, as demonstrated by the partnership that includes medical teams, user associations, and teams from the humanities, social sciences, and neurosciences. The difficulties to be solved depend on the control mode, which is why two lines of research are being developed.
The first axis, assisted teleoperation, focuses on operator assistance in teleoperation situations, acting on the environment via the robot. The second axis, assisted autonomy, seeks to make the robot’s autonomy more reliable by cooperating with an intelligent environment equipped with sensors, actuators, and, more generally, communicating objects (ambient intelligence). This activity has required us to broaden our field of expertise to include distributed and ubiquitous computing.
For more information
Team projects
Consult the projects on which the team is working, as well as completed projects.
Team publications
The team regularly publishes articles, theses, and other documents related to its research.
Team profile
Consult the projects on which the team is working, as well as completed projects.
Upload Team profile →