Interactive co-construction to study dynamical collaborative situations.

Pierre de Loor (1, 2) , Romain Bénard (1) , Cyril Bossard (1)

1 LISYC - Laboratoire d'Informatique des Systèmes Complexes
2 CERV - Centre Européen de Réalité Virtuelle

The purpose of this paper is to present the principle of our framework CoPeFoot dedicated to the study of dynamic and collaborative situations. This research work aims to instate learning by the co-construction of such situations. The article starts by recalling constraints induced by such situations. Next, it introduces interactive co-construction assumption and their implementation in CoPeFoot. In fact, this implementation is based on two steps in CoPeFoot: firstly, machine learning for behavior modeling, using imitation of real users and secondly, refining this behavior by using interaction between the user and the simulation, enhanced by additional information called augmented virtuality. In order to do that, CoPeFoot lies on context base reasoning which is presented. The article ends by a first evaluation of this work.

Mots clés: Co-construction; Context; Case Base Reasoning; Dynamical Situations; Team Sport.; Team Sport

Domaines: Informatique; Système multi-agents [cs.MA]; Intelligence artificielle [cs.AI]; Synthèse d'image et réalité virtuelle [cs.GR]

Pierre de Loor, Romain Bénard, Cyril Bossard. Interactive co-construction to study dynamical collaborative situations.. Proceedings in the International Conference on Virtual Reality, 2008, France. pp.85-91. ⟨hal-00610922⟩