Design and implementation of virtual haptic games for robotic rehabilitation

The following report is concerned with the design and implementation of rotational motion of an object in a virtual environment with appropriate haptic feedback. That is a user of the bimanual system can intuitively control the object in the virtual environment and that haptic feedback gives said user a proprioceptive stimulation creating a strong sense of realism from the experience. This research is built upon previous research conducted in robotics research centre two at the university in producing a bimanual haptic device capable of realistically emulating the motions of the wrist. As such this report will first outline the aforementioned research and endeavours to explain the methods that have been used to create the pre-existing software – the god-object and friction cone algorithms as well some of the existing methods that have not been implemented. Then the report explains the modifications the modifications made to include object rotation in design as well as the reasoning behind those modifications. The new software design was then put to five test subjects to observe and quantify its robustness and was evaluated accordingly. The experiment featured twenty five distinct levels that each of the subjects had a single attempt to complete wherein they were to manipulate a cuboid from one location to a target location and then back to the origin again. The target location would always be the same however it could take on one of five possible orientations each of which occurring five times each throughout the experiment. The whole experimenting process successfully exemplified how the software can be exploited for rehabilitative processes. The results have highlighted the bilateral coupling nature of the arms through the force analysis. The results have also shown that there is a learning process involved in the experiment where most encountered difficulties in the first half of the experiment which steadily declined as the experiment continued. The average time to complete the experiment was 8 minutes and 3 seconds meaning the average time to complete each level (assuming no errors) was 19.34 seconds. This is a fair time to complete the objective of each level though it is noted that the average level success rate is 80% and after treating the failed levels as non-existent the average level completion time is 24.19 seconds. This information all works together to communicate the software under scrutiny is a reasonable approximation to reality in the emulation of grasping and rotating objects.Bachelor of Engineering (Mechanical Engineering