Towards Artificial Intelligence Driven Immersive Environments in Virtual Reality for Industrial Applications

Natural interaction plays a significant role in the credibility of the virtual reality (VR) environments; unwanted or unexpected user interaction with the objects in the environment will negatively affect the immersion level for the user. Users have different level of skills when it comes to using VR. Therefore, handcrafting rule-based robust behaviour interaction that adapts to varying abilities of users is an ongoing challenge. In this study, the potential of using a data-driven method allowing researchers to gain insight into user behavioural data is investigated. A VR data replay and annotation method that allows for the analysis and classification of VR collected data through a graphical user interface is introduced. This method is applied to data collected from a VR lab study, including users with different skills in VR. Finally, the system is used to identify unwanted user interaction and Machine learning methods are investigated as an alternative for user interaction classification.