Spatial knowledge acquisition in a virtual environment

Virtual environments (VE), scaled to emulate natural surroundings, are often assumed to provide accurate spatial knowledge. That assumption was tested by training participants on routes in a spatially complex virtual building. Levels of interactive experience were manipulated by instructional technique and by using a stereoscopic helmet-mounted display (HMD) with or without head-tracking capability. Participants in the "exploratory" group were instructed to navigate by searching for landmarks. Participants in the "restrictive" group were given left-right directions to follow the route. Exploratory instruction resulted in superior route knowledge, despite adverse effects of simulator sickness. Unexpectedly, configurational knowledge was best for exploratory participants without coupled head-tracking, yet it was best for restricted participants with coupled head-tracking. It appears that rapid gaze shifts made by exploratory participants with coupled head-tracking disrupted computer-generated scene updates and interfered with the acquisition of configurational knowledge. Theoretically, with smooth and continuous computer updates, coupled head-tracking for exploratory participants should at least improve configurational knowledge by as much as it did for the restrictive participants, making it by far the superior condition for training. Such gains in configurational knowledge should allow those trained in a VE equipped with coupled head-tracking and given exploratory instruction, to best follow the prescribed route and to be better prepared to take an alternative route if necessary.