ONLINE ASSESSMENT OF TRAINING IN VIRTUAL REALITY SIMULATORS BASED ON GENERAL BAYESIAN NETWORKS

Abstract

Virtual reality systems for training provide
significant benefits over other methods, mainly in critical
procedures. The evaluation of simulations is necessary to
assess the training quality and provide some feedback about
the user performance. An online evaluator for training
simulators based on virtual reality must have a low
complexity algorithm to do not compromise the performance
of the simulator. Several approaches to perform that
evaluation for online or offline training simulators based on
virtual reality have been proposed. We propose a new
approach for an online evaluation based on a General
Bayesian Network. This network was used for modeling and
classification of user’s skills in pre-defined classes of
performance. This method provides the use of continuous
variables without lost of information. Then, it solves the
problem of low complexity in online evaluators without
compromise performance of the simulator.

Index Terms ⎯ General Bayesian networks, virtual reality,
assessment of training, online methods.