Exploring visuo-haptic augmented reality for training

Computer based training has been intensively studied in the last decades and its application in different domains became highly popular. Current systems, however, only provide a fully simulated environment in which the training takes place. Their application is therefore limited by the complexity of the simulated scene or by the capability of the underlying hardware. A concept which can be used to alleviate these problems is augmented reality (AR). Here, a computer generated environment is combined with the real one. While the definition of AR includes augmentation of several modalities, the term still mainly refers to a mere visual augmentation. A visuo-haptic augmented reality therefore extends the basic concept by providing additional haptic sensations from the virtual environment to the user. This allows a user to not only see the virtual world, but in addition to touch it, making more natural interactions possible. In this thesis we investigated visuo-haptic AR and its feasibility to be used in training frameworks. We focused on the fidelity of the AR illusion and the performed research can be divided into different branches: The visuo-haptic augmented reality system itself, the perception within the environment, and the application in different demonstrator systems. Main contributions are the accurate calibration of the haptic interface, the development of a flexible distributed visuo-haptic AR framework, the improvement of visual depth perception by incorporating dynamic shadows and correct handling of occlusions, and an investigation of visuo-haptic perception of compliant objects in AR systems.