Experimental evaluation of interaction design in virtual reality
Desktop Virtual Reality (VR) is a simple and affordable way to implement VR technology into an organisation. With PC technology developing at a phenomenal pace fast processor speeds enable the relatively easy development of visually impressive Virtual Environments (VEs) that can be used with familiar desktop PCs for novice and expert end users alike. A need had consequently evolved to ensure that VE development is structured so that VEs can be visually impressive, usable and effective for their purpose. Interaction between the user and the VE is a distinguishing feature of VR but the importance of interaction on the effectiveness of the VE has been little explored, in particular how to measure that effectiveness with a view to providing guidance to VE developers in this case for training applications using the familiar and affordable desktop medium. The use of VR as a training tool has been widely investigated and implemented in both research and industry. Through experimentation this thesis reviews the design of effective interaction, primarily with the design of selection hotspots (cued objects within the VE designed to prompt the user to select that object) and the importance of implementing task guided interaction into the user’s experience with the VR system. Five experiments were performed to examine the appropriate design of selection hotspots and the importance on the inclusion of a task to the effectiveness of desktop VR training. The initial experiment examined the importance of the user's ability to select within the VE, control their own navigation and the influence of visual realism on the VE’s effectiveness as a training tool. The second experiment explored the importance of the user performing a task on the VE's effectiveness and the effectiveness of various selection hotspot cue designs. The third experiment examined influencing factors on the recall of non-task related aspects of the VE. Experiment four examined the effectiveness of selection hotspot cues when they are no longer congruous to the surrounding VE context and the final experiment investigated if participants perceived and recognised the cued objects or were merely responding to the cue and the influence of the inclusion of cues and their design. Effectiveness was measured using the recall of aspects of the VE by the user and measures of usability, presence and enjoyment. Main findings were that the use of the same incongruous interaction hot spot cues throughout the VE to prompt the selection of specific points within the VE were most effective and using task directed interaction improved task related recall but significantly reduced selection within the VE. Selection significantly increased recall when in a non-task directed VE. With the application of these findings it is possible that designers can produce more effective VEs for their purpose, in this context as a training VE on a desktop VE system.