Enhancing the Prediction of Artificial Lighting Control Behavior Using Virtual Reality (VR): A Pilot Study

Abstract

Virtual reality (VR) has gained reputation in many research areas. It has been increasingly employed in the area of human behavior study. For the study of occupant behavior in buildings, VR is capable of allowing researchers to study occupant behavior in non-existing or future buildings to understand and refine building functions to fulfill occupants’ satisfaction. However, doing research in VR has some limitations specially the sample size. In general, VR experiment generates a fewer number of samples compared to data from sensors a real environment (in-situ). Having a small sample size, sometimes VR data are insufficient to accurately perform further analyses. To overcome this problem, the authors utilize the potential of the Hidden Morkov Model (HMM) Baum-Welch algorithm to statistically learn the sequences and outcomes of VR data and repopulate synthetic data that are good enough for applications. The outcome of repopulated data was evaluated and compared with VR data and showed that HMM precisely estimated and generated the additional VR data.