The effectiveness of training in virtual environments

The research presented in this thesis explores the use of consumer virtual reality technology for training, comparing its validity to more traditional training formats. The need to evaluate the effectiveness of training in virtual environments is critical as a wider audience gains access to an array of emerging virtual reality consumer devices. Training is an obvious use case for these devices. This is motivated by the well-known success of domain-specific training simulators, the ability to train in safe, controlled environments and the potential to launch training programs when the physical components required to complete a task are not readily available. In this thesis, we present four user studies that aim to compare the effectiveness of systems with varying levels of immersion for learning transfer of several tasks, ranging from object location spatial memory to more complex assembly procedures. For every study, evaluation of the effectiveness of training took place in a real-world, physical environment. The first two studies compare geometric and self-motion models in describing human spatial memory through scale distortions of real and virtual environments. The third study examines the effect of level of immersion, self-avatar and environmental fidelity on object location memory in real and virtual environments. The fourth study compares the effectiveness of physical training and virtual training for teaching a bimanual assembly task. Results highlight the validity of virtual environments for training. The overall conclusion is that virtual training can yield a resulting performance that is superior to other, more traditional training formats. Combined, the outcomes of each of the user studies motivate further study of consumer virtual reality systems in training and suggest considerations for the design of such virtual environments