Motion Generation and Planning System for a Virtual Reality Motion Simulator: Development, Integration, and Analysis

In the past five years, the advent of virtual reality devices has significantly influenced research in the field of immersion in a virtual world. In addition to the visual input, the motion cues play a vital role in the sense of presence and the factor of engagement in a virtual environment. This thesis aims to develop a motion generation and planning system for the SP7 motion simulator. SP7 is a parallel robotic manipulator in a 6RSS-R configuration. The motion generation system must be able to produce accurate motion data that matches the visual and audio signals. In this research, two different system workflows have been developed, the first for creating custom visual, audio, and motion cues, while the second for extracting the required motion data from an existing game or simulation. Motion data from the motion generation system are not bounded, while motion simulator movements are limited. The motion planning system commonly known as the motion cueing algorithm is used to create an effective illusion within the limited capabilities of the motion platform. Appropriate and effective motion cues could be achieved by a proper understanding of the perception of human motion, in particular the functioning of the vestibular system. A classical motion cueing has been developed using the model of the semi-circular canal and otoliths. A procedural implementation of the motion cueing algorithm has been described in this thesis. We have integrated all components together to make this robotic mechanism into a VR motion simulator. In general, the performance of the motion simulator is measured by the quality of the motion perceived on the platform by the user. As a result, a novel methodology for the systematic subjective evaluation of the SP7 with a pool of juries was developed to check the quality of motion perception. Based on the results of the evaluation, key issues related to the current configuration of the SP7 have been identified. Minor issues were rectified on the flow, so they were not extensively reported in this thesis. Two major issues have been addressed extensively, namely the parameter tuning of the motion cueing algorithm and the motion compensation of the visual signal in virtual reality devices. The first issue was resolved by developing a tuning strategy with an abstraction layer concept derived from the outcome of the novel technique for the objective assessment of the motion cueing algorithm. The origin of the second problem was found to be a calibration problem of the Vive lighthouse tracking system. So, a thorough experimental study was performed to obtain the optimal calibrated environment. This was achieved by benchmarking the dynamic position tracking performance of the Vive lighthouse tracking system using an industrial serial robot as a ground truth system. With the resolution of the identified issues, a general-purpose virtual reality motion simulator has been developed that is capable of creating custom visual, audio, and motion cues and of executing motion planning for a robotic manipulator with a human motion perception constraint.