Abstract
Power wheelchairs (PWCs) can improve users’ quality of life by enabling them to participate in the activities of daily living, decreasing their dependence on human assistance. PWC users are faced with restricted environments, with limited space to manoeuvre, and are therefore vulnerable to collisions and injuries. To use a PWC effectively and safely, individuals must undertake training and assessment of their competency. There is significant potential for the use of virtual reality in the training and assessment of PWC users.
To date, there is no standard tool available for PWC assessment and training. Rather, clinics use their own observation measurer and assessment is often largely based on guesswork. Several simulators have been developed to help the training of PWC users, yet the study of virtual assessment is an under-researched area. In fact, most simulators offer only very limited functionality and rely solely on client-centric information. For the development of a useful simulator, it is important to identify and evaluate interface factors affecting perception, behaviour, experience, and driving performance from both the user’s and clinician’s perspectives.
In this thesis, issues with current PWC simulators were identified and investigated, with the intention of providing a suitable research platform for the advancement of bringing PWC simulator into clinical use. The aspects investigated include the interaction device, perception and behaviour, and virtual assessment. Three systems were developed to test each of these areas by incorporating theories and techniques from computer science and human-computer interaction.
The first experiment answered the question, “which input devices are necessary and appropriate, and which virtual input device representations can and should be implemented for PWC simulation?” A proprietary PWC joystick was compared to a standard gaming joystick, and driving performance and experience were measured. Four experimental conditions (comprising two virtual input modalities and their two real-world counterparts) were studied. The findings suggest that performance is enhanced when the PWC joystick is represented and that the gaming joystick is adequate for PWC simulation.
The second study investigated the question, “how do immersion factors influence behaviour, perception and sense of presence when navigating a PWC simulator?” The evaluated immersion factors include display type (head mounted display vs. monitor), field of view (changeable vs. static), and avatar presence (present vs. absent). User perception (explicit judgement of doorframe passability) and embedded behaviour (implicit measure of gap passability) were measured, based on the user’s decisions during the experiment. The results show that all three factors affect the user’s sense of presence. The display type affected both perceptual and behavioural measures, whereas field of view only affected behavioural measures.
The final experiment explored the question, “how accurately can clinicians assess driving tasks in the virtual environment compared to the real world?” This study evaluated the effect of three observational techniques (viewpoints) on clinician assessment of PWC driving tasks. In addition, perceived ease of use, confidence level, and sense of presence were also examined. Observational techniques include walk, orbit, and standard viewpoints. The findings of this study suggest that clinicians could make accurate judgments and experience a high confidence level when they were able to walk or orbit the viewpoint. The results from all experiments provide general design guidelines for future virtual reality applications, in particular, PWC simulator design.