RoboGroove: creating fluid motion for dancing robotic arms

Amit Rogel, Richard Savery, Ning Yang, Gil Weinberg

    Research output: Chapter in Book/Report/Conference proceedingConference proceeding contributionpeer-review

    10 Citations (Scopus)
    125 Downloads (Pure)

    Abstract

    Robotic motion has been studied for many purposes, such as effective fast movements, communicative gestures, and obstacle avoidance. Through this study, we are able to improve the perceived expressivity of a robot performing a task by generating trajectories. So far, robots have been very rigid in their movements, making them feel more robotic and less human. The concept of follow through, and smooth movement, can be used to increase animacy for a robot which leads to a more lifelike performance. In this paper, we describe the use of follow through to improve dances and grooves that can match the elegance of human dancers. We created two techniques using a non-humanoid robotic arm, to simulate this for a robotic dancer. Our first technique uses forward kinematics with trajectories that are generated based on a dancer moving to a beat. We mapped various movements of a human dancer to a set of joints on a robotic arm to generate the dancing trajectory. This technique allows a robot to dance in real-time with a human dancer, and also create its own smooth trajectory. The time delay that each human body part uses is implemented as time delay in the robot movements. The second method uses impedance control with varied damping parameters to create follow through. Robotic joints with low damping can passively move in response to other movements in a robot. The arm had high damping for any active moving joints, while the rest of the arm would passively react to this excitation; similar to how a human body responds to our own movement. These two methods were compared in a study where the robot would dance to a beat and users would qualitatively and quantitatively rate the robotic movement. The results of the survey showed that both methods provided an increase in animacy and anthropomorphism of a robot dancing to a beat. Impedance control had the highest rating for animacy, anthropomorphism, full body, and individual body rating. The use of impedance can provide a simple way for a robot to dance like a human, without a change to the primary trajectory.
    Original languageEnglish
    Title of host publicationMOCO '22: Proceedings of 2022 8th International Conference on Movement and Computing
    Place of PublicationNew York
    PublisherAssociation for Computing Machinery (ACM)
    Number of pages9
    ISBN (Electronic)9781450387163
    DOIs
    Publication statusPublished - Jun 2022
    EventInternational Conference on Movement and Computing (8th : 2022) - Chicago, United States
    Duration: 22 Jun 202224 Jun 2022

    Conference

    ConferenceInternational Conference on Movement and Computing (8th : 2022)
    Abbreviated titleMOCO '22
    Country/TerritoryUnited States
    CityChicago
    Period22/06/2224/06/22

    Bibliographical note

    Version archived for private and non-commercial use with the permission of the author/s and according to publisher conditions. For further rights please contact the publisher.

    Keywords

    • fluid motion
    • follow through
    • groove
    • impedance control
    • robot dancing
    • robot perception
    • robotics

    Fingerprint

    Dive into the research topics of 'RoboGroove: creating fluid motion for dancing robotic arms'. Together they form a unique fingerprint.

    Cite this