Semantic modelling as a support for natural interaction within a virtual environment
Abstract
Efficient situated learning in virtual environment requires the trainee to be immerged psychologically. This can be done through natural interaction,
realistic environment and high freedom of action. However, it requires informing all possible actions and behaviours. As virtual reality applications for
training are addressing more complex situations, such elicitation become impossible. It also requires deduction of intention from the user. We propose a semantic representation of the world along with intention recognition techniques to enable a natural interaction without the need for such extensive descriptions