Tool-Use Model Considering Tool Selection by a Robot Using Deep Learning

Namiko Saito, Kitae Kim, Shingo Murata, Tetsuya Ogata, Shigeki Sugano

    研究成果: Conference contribution

    抄録

    We propose a tool-use model that can select tools that require neither labeling nor modeling of the environment and actions. With this model, a robot can choose a tool by itself and perform the operation that matches a human command and the environmental situation. To realize this, we use deep learning to train sensory motor data recorded during tool selection and tool use as experienced by a robot. The experience includes two types of selection, namely according to function and according to size, thereby allowing the robot to handle both situations. For evaluation, the robot is required to generate motion either in an untrained situation or using an untrained tool. We confirm that the robot can choose and use a tool that is suitable for achieving the target task.

    元の言語English
    ホスト出版物のタイトル2018 IEEE-RAS 18th International Conference on Humanoid Robots, Humanoids 2018
    出版者IEEE Computer Society
    ページ814-819
    ページ数6
    ISBN(電子版)9781538672839
    DOI
    出版物ステータスPublished - 2019 1 23
    イベント18th IEEE-RAS International Conference on Humanoid Robots, Humanoids 2018 - Beijing, China
    継続期間: 2018 11 62018 11 9

    出版物シリーズ

    名前IEEE-RAS International Conference on Humanoid Robots
    2018-November
    ISSN(印刷物)2164-0572
    ISSN(電子版)2164-0580

    Conference

    Conference18th IEEE-RAS International Conference on Humanoid Robots, Humanoids 2018
    China
    Beijing
    期間18/11/618/11/9

    Fingerprint

    Robots
    Deep learning
    Labeling

    ASJC Scopus subject areas

    • Artificial Intelligence
    • Computer Vision and Pattern Recognition
    • Hardware and Architecture
    • Human-Computer Interaction
    • Electrical and Electronic Engineering

    これを引用

    Saito, N., Kim, K., Murata, S., Ogata, T., & Sugano, S. (2019). Tool-Use Model Considering Tool Selection by a Robot Using Deep Learning. : 2018 IEEE-RAS 18th International Conference on Humanoid Robots, Humanoids 2018 (pp. 814-819). [8625048] (IEEE-RAS International Conference on Humanoid Robots; 巻数 2018-November). IEEE Computer Society. https://doi.org/10.1109/HUMANOIDS.2018.8625048

    Tool-Use Model Considering Tool Selection by a Robot Using Deep Learning. / Saito, Namiko; Kim, Kitae; Murata, Shingo; Ogata, Tetsuya; Sugano, Shigeki.

    2018 IEEE-RAS 18th International Conference on Humanoid Robots, Humanoids 2018. IEEE Computer Society, 2019. p. 814-819 8625048 (IEEE-RAS International Conference on Humanoid Robots; 巻 2018-November).

    研究成果: Conference contribution

    Saito, N, Kim, K, Murata, S, Ogata, T & Sugano, S 2019, Tool-Use Model Considering Tool Selection by a Robot Using Deep Learning. : 2018 IEEE-RAS 18th International Conference on Humanoid Robots, Humanoids 2018., 8625048, IEEE-RAS International Conference on Humanoid Robots, 巻. 2018-November, IEEE Computer Society, pp. 814-819, 18th IEEE-RAS International Conference on Humanoid Robots, Humanoids 2018, Beijing, China, 18/11/6. https://doi.org/10.1109/HUMANOIDS.2018.8625048
    Saito N, Kim K, Murata S, Ogata T, Sugano S. Tool-Use Model Considering Tool Selection by a Robot Using Deep Learning. : 2018 IEEE-RAS 18th International Conference on Humanoid Robots, Humanoids 2018. IEEE Computer Society. 2019. p. 814-819. 8625048. (IEEE-RAS International Conference on Humanoid Robots). https://doi.org/10.1109/HUMANOIDS.2018.8625048
    Saito, Namiko ; Kim, Kitae ; Murata, Shingo ; Ogata, Tetsuya ; Sugano, Shigeki. / Tool-Use Model Considering Tool Selection by a Robot Using Deep Learning. 2018 IEEE-RAS 18th International Conference on Humanoid Robots, Humanoids 2018. IEEE Computer Society, 2019. pp. 814-819 (IEEE-RAS International Conference on Humanoid Robots).
    @inproceedings{50f1efec8bf94e06b679621557c90e84,
    title = "Tool-Use Model Considering Tool Selection by a Robot Using Deep Learning",
    abstract = "We propose a tool-use model that can select tools that require neither labeling nor modeling of the environment and actions. With this model, a robot can choose a tool by itself and perform the operation that matches a human command and the environmental situation. To realize this, we use deep learning to train sensory motor data recorded during tool selection and tool use as experienced by a robot. The experience includes two types of selection, namely according to function and according to size, thereby allowing the robot to handle both situations. For evaluation, the robot is required to generate motion either in an untrained situation or using an untrained tool. We confirm that the robot can choose and use a tool that is suitable for achieving the target task.",
    author = "Namiko Saito and Kitae Kim and Shingo Murata and Tetsuya Ogata and Shigeki Sugano",
    year = "2019",
    month = "1",
    day = "23",
    doi = "10.1109/HUMANOIDS.2018.8625048",
    language = "English",
    series = "IEEE-RAS International Conference on Humanoid Robots",
    publisher = "IEEE Computer Society",
    pages = "814--819",
    booktitle = "2018 IEEE-RAS 18th International Conference on Humanoid Robots, Humanoids 2018",

    }

    TY - GEN

    T1 - Tool-Use Model Considering Tool Selection by a Robot Using Deep Learning

    AU - Saito, Namiko

    AU - Kim, Kitae

    AU - Murata, Shingo

    AU - Ogata, Tetsuya

    AU - Sugano, Shigeki

    PY - 2019/1/23

    Y1 - 2019/1/23

    N2 - We propose a tool-use model that can select tools that require neither labeling nor modeling of the environment and actions. With this model, a robot can choose a tool by itself and perform the operation that matches a human command and the environmental situation. To realize this, we use deep learning to train sensory motor data recorded during tool selection and tool use as experienced by a robot. The experience includes two types of selection, namely according to function and according to size, thereby allowing the robot to handle both situations. For evaluation, the robot is required to generate motion either in an untrained situation or using an untrained tool. We confirm that the robot can choose and use a tool that is suitable for achieving the target task.

    AB - We propose a tool-use model that can select tools that require neither labeling nor modeling of the environment and actions. With this model, a robot can choose a tool by itself and perform the operation that matches a human command and the environmental situation. To realize this, we use deep learning to train sensory motor data recorded during tool selection and tool use as experienced by a robot. The experience includes two types of selection, namely according to function and according to size, thereby allowing the robot to handle both situations. For evaluation, the robot is required to generate motion either in an untrained situation or using an untrained tool. We confirm that the robot can choose and use a tool that is suitable for achieving the target task.

    UR - http://www.scopus.com/inward/record.url?scp=85062265463&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85062265463&partnerID=8YFLogxK

    U2 - 10.1109/HUMANOIDS.2018.8625048

    DO - 10.1109/HUMANOIDS.2018.8625048

    M3 - Conference contribution

    AN - SCOPUS:85062265463

    T3 - IEEE-RAS International Conference on Humanoid Robots

    SP - 814

    EP - 819

    BT - 2018 IEEE-RAS 18th International Conference on Humanoid Robots, Humanoids 2018

    PB - IEEE Computer Society

    ER -