PUB - Publikationen an der Universität Bielefeld

Nowadays, virtual reality (VR) technology is becoming increasingly realistic, including the possibility to touch, feel, and interact with the virtual world by using haptic gloves. Being able to create a tangible virtual world through programming opens up new possibilities, even in more serious contexts, such as haptic performance evaluation during stroke rehabilitation. Revising traditional tests is necessary because they are based on coarse-grained measured task execution or movement observations. Creating assessments with VR and haptic gloves offer to measure detailed motion data and allows flexible haptic task implementations. By adding game elements, and thus long-term motivation, the evaluation framework can become a training tool. Therefore, the aim of this thesis was to create a VR serious game framework with haptic gloves for haptic performance evaluation.
The serious part consisted of VR haptic scenarios built upon related work reporting various diverging haptic sub-skills, as well as relevant task execution and movement parameters. The game part was implemented as a tower defense game containing a genetic algorithm for difficulty adaptation. The algorithm and the game experience were shown to be successful with two studies and a data evaluation.
The haptic performance evaluation part was approached exemplarily with a haptic rotation task in reality. A pilot study and three experiments were conducted to find a standardized procedure resulting in a suitable task execution parameter distribution. Incidental findings were that people tend to turn over the target, rotary knob shapes and the number of fingers involved affect the rotation accuracy, and start orientations affect the precision. Thereafter, regression models identified which movement parameters are connected to good task performance within a healthy sample.
For the haptic glove part, two conducted experiments showed that task execution and movement parameters diverge between VR and reality. Based on the assumption that the haptic glove’s under-determined joint angle measurement is responsible for the disparities, a haptic glove augmentation framework was developed. It consisted of additional hardware to measure missing joints, a hand model configuration tool (HMCT), and software implementing forward and inverse kinematics for hand posture calculation. Based on a performed study, the HMCT provided accurate hand models. A study revealed that the hand posture calculation was superior to the original heuristic, but there was still room for improvement.
For future work, the haptic glove augmentation framework needs to be improved, and the VR scenes should be turned into standardized tests. Overall, the main contributions in the field of haptic performance evaluation were to develop an extendable open-source framework, propose a more systematic approach, to reveal the benefits of detailed motion data recording, and to identify critical challenges in the use of haptic gloves and VR.