Human-robot communication using multiple recurrent neural networks

Yoshihiro Sakamoto, Tetsuya Ogata, Shigeki Sugano

    研究成果: Conference contribution

    抄録

    On the methodology of robotic design from the traditional view of communication which is assumed as a symbol process, robots are forced to confront the symbol grounding problem. However, if communication is assumed as the analog dynamics and robots are driven by it, robots can avoid the problem and be situated in the environment and to other agents. In this paper we will introduce a new communication system constructed from the view of dynamical systems to achieve the situatedness. This system is that there is a robot in a virtual environment and the control of the robot is shared by human operation using a joystick and a robot controller. As the controller, we adopt multiple recurrent neural networks (MRNN) which are able to cope with complex environments and broad communication that single recurrent net cannot cope with. We conduct two experiments in order to evaluate the effectiveness of MRNN to a low level communication task such as nonverbal interaction. First, we examine the effect of the number of RNNs contained in MRNN. Second, we examine the effect of the context dependency of MRNN. These experiments show the capability of MRNN as a new-type controller of communication robot.

    元の言語English
    ホスト出版物のタイトル2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
    ページ1574-1579
    ページ数6
    2
    出版物ステータスPublished - 2004
    イベント2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) - Sendai
    継続期間: 2004 9 282004 10 2

    Other

    Other2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
    Sendai
    期間04/9/2804/10/2

    Fingerprint

    Recurrent neural networks
    Robots
    Communication
    Controllers
    Electric grounding
    Virtual reality
    Communication systems
    Dynamical systems
    Robotics
    Experiments

    ASJC Scopus subject areas

    • Engineering(all)

    これを引用

    Sakamoto, Y., Ogata, T., & Sugano, S. (2004). Human-robot communication using multiple recurrent neural networks. : 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (巻 2, pp. 1574-1579)

    Human-robot communication using multiple recurrent neural networks. / Sakamoto, Yoshihiro; Ogata, Tetsuya; Sugano, Shigeki.

    2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). 巻 2 2004. p. 1574-1579.

    研究成果: Conference contribution

    Sakamoto, Y, Ogata, T & Sugano, S 2004, Human-robot communication using multiple recurrent neural networks. : 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). 巻. 2, pp. 1574-1579, 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Sendai, 04/9/28.
    Sakamoto Y, Ogata T, Sugano S. Human-robot communication using multiple recurrent neural networks. : 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). 巻 2. 2004. p. 1574-1579
    Sakamoto, Yoshihiro ; Ogata, Tetsuya ; Sugano, Shigeki. / Human-robot communication using multiple recurrent neural networks. 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). 巻 2 2004. pp. 1574-1579
    @inproceedings{9fa58b0351b74785bbd4cd5798399b15,
    title = "Human-robot communication using multiple recurrent neural networks",
    abstract = "On the methodology of robotic design from the traditional view of communication which is assumed as a symbol process, robots are forced to confront the symbol grounding problem. However, if communication is assumed as the analog dynamics and robots are driven by it, robots can avoid the problem and be situated in the environment and to other agents. In this paper we will introduce a new communication system constructed from the view of dynamical systems to achieve the situatedness. This system is that there is a robot in a virtual environment and the control of the robot is shared by human operation using a joystick and a robot controller. As the controller, we adopt multiple recurrent neural networks (MRNN) which are able to cope with complex environments and broad communication that single recurrent net cannot cope with. We conduct two experiments in order to evaluate the effectiveness of MRNN to a low level communication task such as nonverbal interaction. First, we examine the effect of the number of RNNs contained in MRNN. Second, we examine the effect of the context dependency of MRNN. These experiments show the capability of MRNN as a new-type controller of communication robot.",
    author = "Yoshihiro Sakamoto and Tetsuya Ogata and Shigeki Sugano",
    year = "2004",
    language = "English",
    isbn = "0780384636",
    volume = "2",
    pages = "1574--1579",
    booktitle = "2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)",

    }

    TY - GEN

    T1 - Human-robot communication using multiple recurrent neural networks

    AU - Sakamoto, Yoshihiro

    AU - Ogata, Tetsuya

    AU - Sugano, Shigeki

    PY - 2004

    Y1 - 2004

    N2 - On the methodology of robotic design from the traditional view of communication which is assumed as a symbol process, robots are forced to confront the symbol grounding problem. However, if communication is assumed as the analog dynamics and robots are driven by it, robots can avoid the problem and be situated in the environment and to other agents. In this paper we will introduce a new communication system constructed from the view of dynamical systems to achieve the situatedness. This system is that there is a robot in a virtual environment and the control of the robot is shared by human operation using a joystick and a robot controller. As the controller, we adopt multiple recurrent neural networks (MRNN) which are able to cope with complex environments and broad communication that single recurrent net cannot cope with. We conduct two experiments in order to evaluate the effectiveness of MRNN to a low level communication task such as nonverbal interaction. First, we examine the effect of the number of RNNs contained in MRNN. Second, we examine the effect of the context dependency of MRNN. These experiments show the capability of MRNN as a new-type controller of communication robot.

    AB - On the methodology of robotic design from the traditional view of communication which is assumed as a symbol process, robots are forced to confront the symbol grounding problem. However, if communication is assumed as the analog dynamics and robots are driven by it, robots can avoid the problem and be situated in the environment and to other agents. In this paper we will introduce a new communication system constructed from the view of dynamical systems to achieve the situatedness. This system is that there is a robot in a virtual environment and the control of the robot is shared by human operation using a joystick and a robot controller. As the controller, we adopt multiple recurrent neural networks (MRNN) which are able to cope with complex environments and broad communication that single recurrent net cannot cope with. We conduct two experiments in order to evaluate the effectiveness of MRNN to a low level communication task such as nonverbal interaction. First, we examine the effect of the number of RNNs contained in MRNN. Second, we examine the effect of the context dependency of MRNN. These experiments show the capability of MRNN as a new-type controller of communication robot.

    UR - http://www.scopus.com/inward/record.url?scp=14044275128&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=14044275128&partnerID=8YFLogxK

    M3 - Conference contribution

    AN - SCOPUS:14044275128

    SN - 0780384636

    VL - 2

    SP - 1574

    EP - 1579

    BT - 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

    ER -