Using HMD for Immersive Training of Voice-Based Operation of Small Unmanned Ground Vehicles

Voice recognition systems provide a method of hands-free control of robotic systems that may be helpful in law enforcement or military domains. However, the constraints of the operational environment limit the capabilities of the on-board voice recognition system to a keyword-based command system. To effectively use the system, the users must learn the available commands and practice pronunciation to ensure accurate recognition. Virtual reality simulation provides users the opportunity to train with the voice recognition system and the robot platform in realistic interactive scenarios. Training using virtual reality with a head-mounted display may increase immersion and sense of presence compared to using a keyboard and monitor. A small pilot study compared user experience in the desktop mode and the virtual reality mode of our voice recognition system training tool. Participants controlled a simulated unmanned ground vehicle in two different modes across four different environments. The results revealed no significant differences in simulator sickness, sense of presence, or perceived usability. However, when asked to choose between the desktop mode and the head-mounted display mode, results indicate users' overall preference for the head-mounted display. However, the users also perceive the head-mounted display to be more complex, less consistent, and more difficult to learn to use. The desktop mode was perceived as easier to use and users reported being more confident when using it.