Domitile Lourdeaux

Relevance of an intelligent tutorial agent for virtual reality training systems

Virtual reality (VR) is mostly used as a technical means of immersing the user in a computer-generated situation where the reality is the model to be perfectly recreated. Trainers often provide supportive explanations by using tutorial situations that do not exactly reproduce reality. Therefore, VR systems should also provide such tutorial situations. In order to address this issue, an intelligent tutorial agent has been designed to support the learning process. The agent is based on a cognitive taxonomy of the trainees behaviour in the VR context. For each trainee, the agent provides two kinds of tutorial according to his/her knowledge level. The first one consists of modifying the scenario and the second one consists of guiding the trainee. These strategies can be represented by different fidelity levels of virtual environments (e.g. enhancement, simplification). They can be activated by the trainee, by the trainer or automatically. They are specified off-line and continuously modified on-line. An application of this intelligent agent is being implemented in a VR simulation-based training system for SNCF train drivers. The goal of this application is to teach the procedure to control and handle switches on high speed train tracks.

An industrial approach to design compelling VR and AR experience

The convergence of technologies currently observed in the field of VR, AR, robotics and consumer electronic reinforces the trend of new applications appearing every day. But when transferring knowledge acquired from research to businesses, research laboratories are often at a loss because of a lack of knowledge of the design and integration processes in creating an industrial scale product. In fact, the innovation approaches that take a good idea from the laboratory to a successful industrial product are often little known to researchers. The objective of this paper is to present the results of the work of several research teams that have finalized a working method for researchers and manufacturers that allow them to design virtual or augmented reality systems and enable their users to enjoy “a compelling VR experience”. That approach, called “the I2I method”, present 11 phases from “Establishing technological and competitive intelligence and industrial property” to “Improvements” through the “Definition of the Behavioral Interface, Virtual Environment and Behavioral Software Assistance”. As a result of the experience gained by various research teams, this design approach benefits from contributions from current VR and AR research. Our objective is to validate and continuously move such multidisciplinary design team methods forward.

I Feel What You Feel: Empathy and Placebo Mechanisms for Autonomous Virtual Humans

Computational modeling of emotion, physiology and personality is a major challenge in order to design believable virtual humans. These factors have an impact on both the individual behavior and the collective one. This requires to take into account the empathy phenomenon. Furthermore, in a crisis simulation context where the virtual humans can be contaminated by radiological or chemical substances, empathy may lead to placebo or nocebo effects. Stemming from works in the multiagent systems domain, our virtual human decision process is designed as an autonomous agent. It has been shown that the environment can encapsulate the responsibility of spreading part of the agent state. The agent has two parts, its mind and its body. The mind contains the decision process and is autonomous. The body is influenced by the mind, but controlled by the environment which manages the empathy process. Combined with biased reasoning, favorable personality traits and situational factors, empathy can lead some agents to believe they are contaminated although they are not. We describe these mechanisms and show the results of several experiments.

Activité collective et personnages autonomes : moteur décisionnel reposant sur des relations de confiance

When working in teams, people make mistakes. To train someone in a collaborative virtual environment to adapt to teammates that bahave non optimally, we propose (1) an augmentation of the ACTIVITY-Description Language as well as mechanisms of propagation of constraints that will facilitate agents’ reasoning; and (2) an agent model in which each agent is described through three dimensions (integrity, benevolence, abilities) corresponding to the MDS trust model. Besides each agent has different personal and collective goals and has beliefs about others’ integrity, benevolence and abilities. This agent model is associated to a decision-making system that allows agents to adopt human-like behaviors. In particular, agents take others into account and are able to reason on their beliefs about others both when choosing which goal (collective or individual) to focus on and when selecting a task. We conducted a preliminary evaluation in which participants evaluated the behaviors produced with our system. MOTS-CLÉS : systèmes multi-agents, prise de décision, confiance, activité collective.