A virtual environment for learning to pilot remotely operated vehicles

Remotely operated vehicles (ROVs) are used extensively for underwater searching and salvage, inspection, surveying, scientific exploration and mine countermeasures. ROV pilots must learn to rely on limited data from video and sonar displays and a few other positional indicators to maintain a sense of their vehicle, its tether and its surroundings. Pilot training typically occurs on-the-job, where equipment is placed at risk, controlled learning situations are hard to create, and time for instruction is minimal. The TRANSoM (TRAiNing for remote Sensing and Manipulation) project is developing a training environment that combines two emerging technologies to overcome these limitations. A virtual environment (VE) simulates the ROV and its surroundings, and also has instructional enhancements such as external views of the ROV, directional cues and alternate modes of interaction with the vehicle, e.g. a head-tracked head-mounted display (HMD). An intelligent tutoring system (ITS) monitors student pilot behavior and offers verbal and graphical feedback, a mission review and a performance assessment. Near-transfer experiments comparing the utility of different artificial viewpoints have been completed, with full transfer experiments involving a real ROV to follow.