Virtual Reality and Deep Learning are technologies that can enhance Human Engineering. This paper discusses some experiments using paradigms of neural networks (Machine Learning) vs. Deep Learning to predict the NASA Task Load Index (TLX) from general information and brain waves without answering the questionnaire. The environment of the experiments was a virtual reality-based one using OpenSim to build it and Oculus Rift to navigate it. Electroencephalograms were used to measure the level of engagement of the users. After the tasks were performed, the TLX questionnaire was completed. Resilient Backpropagation and Deep Learning-based architectures were used to create a mapping of brain waves and demographic information to the TLX index. The results are very positive, and the TLX index can be predicted. These results can support the development of real-time assessment systems and the build-up of adaptable/smart user interfaces.
Comparing Backpropagation-based Neural Networks and Deep Learning to Predict the NASA Task Load Index
2 Views
8 Downloads