View Item 
        •   Utrecht University Student Theses Repository Home
        • UU Theses Repository
        • Theses
        • View Item
        •   Utrecht University Student Theses Repository Home
        • UU Theses Repository
        • Theses
        • View Item
        JavaScript is disabled for your browser. Some features of this site may not work without it.

        Browse

        All of UU Student Theses RepositoryBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

        Live Control of a Physically Plausible Virtual Biped, Using the Kinect

        Thumbnail
        View/Open
        ICA-3858936.pdf (2.232Mb)
        Publication date
        2014
        Author
        Zwennes, O.
        Metadata
        Show full item record
        Summary
        With the availability of affordable motion capture solutions, like the Microsoft Kinect, live data-driven control of a virtual character has become more relevant. In order to provide both intuitive control of a virtual biped, as well as grounding that character in the virtual environment, we propose a system for real-time control of a physically plausible virtual biped using the Microsoft Kinect sensor. In order to deal with the varying skeletal dimensions present in Kinect data, we propose an inverse kinematics based real-time data processing method, which avoids positional errors of the limbs in the virtual character, while maintaining the pose present in the Kinect data. In order to allow for arbitrary in-place motions to be performed by the physically plausible character without losing balance, we combine proportional derivative controllers with a simple balance strategy that controls the center of mass of the character by rotating the ankle joints. Experimental results show that with a limited range of balanced in-place motions, our system is able to reproduce these motions with the physics simulated character in close to real-time. Our system achieves this using solely physically plausible internal torques and using the Kinect data stream as its only motion data input.
        URI
        https://studenttheses.uu.nl/handle/20.500.12932/16366
        Collections
        • Theses
        Utrecht university logo