Training in virtual environments: analysis of task appropriateness

Abstract

Because of its ability to furnish complex interactive visual and auditory stimuli, virtual reality has been lauded as a wonder tool for training. Surprisingly, very few articles have been published which critically examine these claims, or which even empirically examine them. The studies which have empirically looked at virtual reality training have produced conflicting results. Just as in any training situation, the tasks to be trained must be decomposed into their component cognitive, perceptual and motor demands, and these demands must be met in the training environment. In short, to create an effective training environment, one must match up the capabilities of the training environment with the demands of the actual task. Virtual environments (VEs) are currently very weak at haptic displays, kinesthetic feedback and vestibular feedback. VEs are strongest in visual information display and head motion feedback. The tasks which are most likely to benefit from training in a virtual environment are tasks which heavily depend on visual information for success. Attempting to train a manual placement task, a task whose performance relies as much on haptic and kinesthetic feedback as on fine visual feedback, will not work. On the other hand, training for a visuospatial perception task matches up well with the characteristics of current VE systems. Tasks such as inspection tasks and navigation tasks take advantage of VEs' strengths for training purposes. To test this logic, we trained a navigation task using a VE.