An Intelligent Virtual Environment for Training with Dynamic Parameters
Lyudmila Steshina, Igor Petukhov, Andrey Glazyrin, Пламена Златева, Dimiter Georgiev Velev · 2020
The paper proposes a new approach for training people in Virtual Reality environment. The key idea is the intellectualization of the Virtual Environment by changing the dynamic parameters, taking into account the performance metrics. The performance metrics include the time spent for the current task and the mistakes made. If the learner makes some errors within a long time interval, then the environment is rebuilt by using the dynamic parameters. The mechanism for the changing the dynamic parameters in the Virtual Reality, based on the fuzzy model with a set of rules and the membership functions, bind the performance metrics and the characteristic for those parameters. At the same time, the complex environment can include hundreds of dynamic parameters and for this case, an optimization of the fuzzy model is needed. For this reason, a genetic algorithm with a single fuzzy rule for changing the dynamic parameters is applied. This approach is used in the development of a training simulator for a forest cutting machine (harvester). The results obtained could be beneficial to modeling of training systems and for improving the performance during the training session.