The goal of the project INGREDIBLE is to propose a set of scientific innovations in the domain of interactions between humans and virtual actors. The project aims at developing a prototype of an autonomous virtual actor capable a dynamic affective bodily coupling with a human whose gesture are analyzed in real time. The main focus is on the dynamic of the coupling of the gestures which is considered to be the key point of the feeling of the “presence of the other” and the “involvement” of the users in the virtual reality systems of the future. Each actor's gesture or variation in the gesture changes the other's and vice versa. Depending of the emotional states of the participants, it can lead to the emergence of coordination, a gesture based dialogue, composed of initiatives and imitations which is singular. It is this singularity that is specific to these “real” human interactions which we want to be translated in the real/virtual space. In particular, our approach will include in a particular manner several observed phenomenon in body interaction : integration of the different channel carrying the expressiveness (body position, quality of the movement, gestures), imitation, postural convergence and synchrony. The project will focus on the body expression of the emotions in situations of interaction. It will also focus on the dynamics of the chain of the gestures. The project will not tackle the linguistic and oral aspects. This choice is justified by a realist estimate of the complexity of the task to be realized in a limited time. The real time gesture based interaction alone is a full-fledged domain which can be separated from the analysis of the speech in a first approach. In order to reach the full wanted behavioral richness , the interaction must be contextualized by the mean of predefined scenarios. The fact the mechanisms of body interaction are generic, through different contexts, allows this separate study to better prepare a future link between oral and body interaction. For the content, the project tends to allow a synergy between several scientific domains : - Study and formalization of the dynamic of the body interaction between humans - Study on the body expression and posture of the emotions - Real time gesture recognition and anticipation - Model of autonomous interactive agents - Computer generation of expressive movements and adaptation in real time The projet INGREDIBLE is also learning toward applications, using : - A generic approach using several animation platforms and several scenarios for the application. - An evaluation of the capacity offered by several peripheral (Kinect, Optitrack, Moven, …) and animation platform (MARC, AréVi/HLIB, SMR). - Scenarios for the use of the prototype coming from different domains (drama, sport, education). - Users proposing new scenarios et allowing the guarantee the accuracy of the needs.

How can we counter the risk of "information flood" caused by the BD (Battlespace Digitization)? The purpose of this project is to provide a range of tools, complementing the traditional Cartographic Information System (CIS), to facilitate and accelerate the understanding of complex tactical situations. The idea is to propose, for every military (logistics, engineering, combat, future manoeuvre ...) or civil (security, victims rescue, evacuation, ...) functional profile, three complementary tools: (1) a visualisation of the common operational picture and ongoing actions; (2) an aid for understanding through an analysis of past events / operational reports offering an automatic synthesis of history; and (3) a collaborative tool for manœuvre conception. This trilogy allows to equip officers and decision makers in their understanding of complex situations, in their briefing work, thus accelerating the decision-making loop within the PC. It will be offered on touch pad to provide an easily deployable tool, mobile with each decision-maker, allowing them to work independently or collaboratively. Moreover each of these bricks is customised depending on the functional profile of the officer (infantry, logistics and engineering). In this project, we propose to enrich the outcome model from the initial ASTRID project, TACTIC, to meet such a matrix of specific needs by developing a range of tools for (1) offering a “function” view (weapon, profession) of the Common Operational Picture (COP); (2) facilitating the understanding of the current situation through an analysis of the causal structure of past events thanks to a new Sense Making technology inspired by Narrative Intelligence research; (3) accelerating the integration of orders and communication within the decision cell by providing an optimised and intelligent order transmission or tactical drawing system. In order to do this, and to increase the TRL of the outcome model from TACTIC , it is necessary to rework some functional components such as the cartographic management, or to enrich the coupling with SWORD in order to take full advantage of the simulated environment, and benefit from a realistic environment for evaluation and validation. MASA Group, a SME, and the laboratory LabSTICC, continue to work together to enrich and improve the outcome model of the initial ASTRID project. In order to offer an evaluation in a realistic environment, demonstrators of tools from the project will be stimulated thanks to a constructive C2 simulation operationally used by the army, SWORD, which will generate reports and realistic accounts in the context of operational scenarios.