Dynamic motion learning for multi-DOF flexible-joint robots using active–passive motor babbling through deep learning

Kuniyuki Takahashi, Tetsuya Ogata, Jun Nakanishi, Gordon Cheng, Shigeki Sugano

    Research output: Contribution to journalArticle

    1 Citation (Scopus)

    Abstract

    This paper proposes a learning strategy for robots with flexible joints having multi-degrees of freedom in order to achieve dynamic motion tasks. In spite of there being several potential benefits of flexible-joint robots such as exploitation of intrinsic dynamics and passive adaptation to environmental changes with mechanical compliance, controlling such robots is challenging because of increased complexity of their dynamics. To achieve dynamic movements, we introduce a two-phase learning framework of the body dynamics of the robot using a recurrent neural network motivated by a deep learning strategy. The proposed methodology comprises a pre-training phase with motor babbling and a fine-tuning phase with additional learning of the target tasks. In the pre-training phase, we consider active and passive exploratory motions for efficient acquisition of body dynamics. In the fine-tuning phase, the learned body dynamics are adjusted for specific tasks. We demonstrate the effectiveness of the proposed methodology in achieving dynamic tasks involving constrained movement requiring interactions with the environment on a simulated robot model and an actual PR2 robot both of which have a compliantly actuated seven degree-of-freedom arm. The results illustrate a reduction in the required number of training iterations for task learning and generalization capabilities for untrained situations.

    Original languageEnglish
    Pages (from-to)1002-1015
    Number of pages14
    JournalAdvanced Robotics
    Volume31
    Issue number18
    DOIs
    Publication statusPublished - 2017 Sep 17

    Fingerprint

    Robots
    Tuning
    Deep learning
    Recurrent neural networks
    Degrees of freedom (mechanics)

    Keywords

    • deep learning
    • dynamic motion learning
    • flexible-joint robot
    • Motor babbling
    • recurrent neural network

    ASJC Scopus subject areas

    • Control and Systems Engineering
    • Software
    • Human-Computer Interaction
    • Hardware and Architecture
    • Computer Science Applications

    Cite this

    Dynamic motion learning for multi-DOF flexible-joint robots using active–passive motor babbling through deep learning. / Takahashi, Kuniyuki; Ogata, Tetsuya; Nakanishi, Jun; Cheng, Gordon; Sugano, Shigeki.

    In: Advanced Robotics, Vol. 31, No. 18, 17.09.2017, p. 1002-1015.

    Research output: Contribution to journalArticle

    @article{b85d665cf1e44bbc85a0b5f852ccfab5,
    title = "Dynamic motion learning for multi-DOF flexible-joint robots using active–passive motor babbling through deep learning",
    abstract = "This paper proposes a learning strategy for robots with flexible joints having multi-degrees of freedom in order to achieve dynamic motion tasks. In spite of there being several potential benefits of flexible-joint robots such as exploitation of intrinsic dynamics and passive adaptation to environmental changes with mechanical compliance, controlling such robots is challenging because of increased complexity of their dynamics. To achieve dynamic movements, we introduce a two-phase learning framework of the body dynamics of the robot using a recurrent neural network motivated by a deep learning strategy. The proposed methodology comprises a pre-training phase with motor babbling and a fine-tuning phase with additional learning of the target tasks. In the pre-training phase, we consider active and passive exploratory motions for efficient acquisition of body dynamics. In the fine-tuning phase, the learned body dynamics are adjusted for specific tasks. We demonstrate the effectiveness of the proposed methodology in achieving dynamic tasks involving constrained movement requiring interactions with the environment on a simulated robot model and an actual PR2 robot both of which have a compliantly actuated seven degree-of-freedom arm. The results illustrate a reduction in the required number of training iterations for task learning and generalization capabilities for untrained situations.",
    keywords = "deep learning, dynamic motion learning, flexible-joint robot, Motor babbling, recurrent neural network",
    author = "Kuniyuki Takahashi and Tetsuya Ogata and Jun Nakanishi and Gordon Cheng and Shigeki Sugano",
    year = "2017",
    month = "9",
    day = "17",
    doi = "10.1080/01691864.2017.1383939",
    language = "English",
    volume = "31",
    pages = "1002--1015",
    journal = "Advanced Robotics",
    issn = "0169-1864",
    publisher = "Taylor and Francis Ltd.",
    number = "18",

    }

    TY - JOUR

    T1 - Dynamic motion learning for multi-DOF flexible-joint robots using active–passive motor babbling through deep learning

    AU - Takahashi, Kuniyuki

    AU - Ogata, Tetsuya

    AU - Nakanishi, Jun

    AU - Cheng, Gordon

    AU - Sugano, Shigeki

    PY - 2017/9/17

    Y1 - 2017/9/17

    N2 - This paper proposes a learning strategy for robots with flexible joints having multi-degrees of freedom in order to achieve dynamic motion tasks. In spite of there being several potential benefits of flexible-joint robots such as exploitation of intrinsic dynamics and passive adaptation to environmental changes with mechanical compliance, controlling such robots is challenging because of increased complexity of their dynamics. To achieve dynamic movements, we introduce a two-phase learning framework of the body dynamics of the robot using a recurrent neural network motivated by a deep learning strategy. The proposed methodology comprises a pre-training phase with motor babbling and a fine-tuning phase with additional learning of the target tasks. In the pre-training phase, we consider active and passive exploratory motions for efficient acquisition of body dynamics. In the fine-tuning phase, the learned body dynamics are adjusted for specific tasks. We demonstrate the effectiveness of the proposed methodology in achieving dynamic tasks involving constrained movement requiring interactions with the environment on a simulated robot model and an actual PR2 robot both of which have a compliantly actuated seven degree-of-freedom arm. The results illustrate a reduction in the required number of training iterations for task learning and generalization capabilities for untrained situations.

    AB - This paper proposes a learning strategy for robots with flexible joints having multi-degrees of freedom in order to achieve dynamic motion tasks. In spite of there being several potential benefits of flexible-joint robots such as exploitation of intrinsic dynamics and passive adaptation to environmental changes with mechanical compliance, controlling such robots is challenging because of increased complexity of their dynamics. To achieve dynamic movements, we introduce a two-phase learning framework of the body dynamics of the robot using a recurrent neural network motivated by a deep learning strategy. The proposed methodology comprises a pre-training phase with motor babbling and a fine-tuning phase with additional learning of the target tasks. In the pre-training phase, we consider active and passive exploratory motions for efficient acquisition of body dynamics. In the fine-tuning phase, the learned body dynamics are adjusted for specific tasks. We demonstrate the effectiveness of the proposed methodology in achieving dynamic tasks involving constrained movement requiring interactions with the environment on a simulated robot model and an actual PR2 robot both of which have a compliantly actuated seven degree-of-freedom arm. The results illustrate a reduction in the required number of training iterations for task learning and generalization capabilities for untrained situations.

    KW - deep learning

    KW - dynamic motion learning

    KW - flexible-joint robot

    KW - Motor babbling

    KW - recurrent neural network

    UR - http://www.scopus.com/inward/record.url?scp=85033230563&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85033230563&partnerID=8YFLogxK

    U2 - 10.1080/01691864.2017.1383939

    DO - 10.1080/01691864.2017.1383939

    M3 - Article

    AN - SCOPUS:85033230563

    VL - 31

    SP - 1002

    EP - 1015

    JO - Advanced Robotics

    JF - Advanced Robotics

    SN - 0169-1864

    IS - 18

    ER -