Sensor prediction and grasp stability evaluation for in-hand manipulation

Kohei Kojima, Takashi Sato, Alexander Schmitz, Hiroaki Arie, Hiroyasu Iwata, Shigeki Sugano

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    9 Citations (Scopus)

    Abstract

    Handling objects with a single hand without dropping the object is challenging for a robot. A possible way to aid the motion planning is the prediction of the sensory results of different motions. Sequences of different movements can be performed as an offline simulation, and using the predicted sensory results, it can be evaluated whether the desired goal is achieved. In particular, the task in this paper is to roll a sphere between the fingertips of the dexterous hand of the humanoid robot TWENDY-ONE. First, a forward model for the prediction of the touch state resulting from the in-hand manipulation is developed. As it is difficult to create such a model analytically, the model is obtained through machine learning. To get real world training data, a dataglove is used to control the robot in a master-slave way. The learned model was able to accurately predict the course of the touch state while performing successful and unsuccessful in-hand manipulations. In a second step, it is shown that this simulated sequence of sensor states can be used as input for a stability assessment model. This model can accurately predict whether a grasp is stable or whether it results in dropping the object. In a final step, a more powerful grasp stability evaluator is introduced, which works for our task regardless of the sphere diameter.

    Original languageEnglish
    Title of host publicationIEEE International Conference on Intelligent Robots and Systems
    Pages2479-2484
    Number of pages6
    DOIs
    Publication statusPublished - 2013
    Event2013 26th IEEE/RSJ International Conference on Intelligent Robots and Systems: New Horizon, IROS 2013 - Tokyo
    Duration: 2013 Nov 32013 Nov 8

    Other

    Other2013 26th IEEE/RSJ International Conference on Intelligent Robots and Systems: New Horizon, IROS 2013
    CityTokyo
    Period13/11/313/11/8

    Fingerprint

    Sensors
    Robots
    End effectors
    Motion planning
    Learning systems

    ASJC Scopus subject areas

    • Control and Systems Engineering
    • Software
    • Computer Vision and Pattern Recognition
    • Computer Science Applications

    Cite this

    Kojima, K., Sato, T., Schmitz, A., Arie, H., Iwata, H., & Sugano, S. (2013). Sensor prediction and grasp stability evaluation for in-hand manipulation. In IEEE International Conference on Intelligent Robots and Systems (pp. 2479-2484). [6696705] https://doi.org/10.1109/IROS.2013.6696705

    Sensor prediction and grasp stability evaluation for in-hand manipulation. / Kojima, Kohei; Sato, Takashi; Schmitz, Alexander; Arie, Hiroaki; Iwata, Hiroyasu; Sugano, Shigeki.

    IEEE International Conference on Intelligent Robots and Systems. 2013. p. 2479-2484 6696705.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Kojima, K, Sato, T, Schmitz, A, Arie, H, Iwata, H & Sugano, S 2013, Sensor prediction and grasp stability evaluation for in-hand manipulation. in IEEE International Conference on Intelligent Robots and Systems., 6696705, pp. 2479-2484, 2013 26th IEEE/RSJ International Conference on Intelligent Robots and Systems: New Horizon, IROS 2013, Tokyo, 13/11/3. https://doi.org/10.1109/IROS.2013.6696705
    Kojima K, Sato T, Schmitz A, Arie H, Iwata H, Sugano S. Sensor prediction and grasp stability evaluation for in-hand manipulation. In IEEE International Conference on Intelligent Robots and Systems. 2013. p. 2479-2484. 6696705 https://doi.org/10.1109/IROS.2013.6696705
    Kojima, Kohei ; Sato, Takashi ; Schmitz, Alexander ; Arie, Hiroaki ; Iwata, Hiroyasu ; Sugano, Shigeki. / Sensor prediction and grasp stability evaluation for in-hand manipulation. IEEE International Conference on Intelligent Robots and Systems. 2013. pp. 2479-2484
    @inproceedings{581d5235199f47a7bacbb8861012e0de,
    title = "Sensor prediction and grasp stability evaluation for in-hand manipulation",
    abstract = "Handling objects with a single hand without dropping the object is challenging for a robot. A possible way to aid the motion planning is the prediction of the sensory results of different motions. Sequences of different movements can be performed as an offline simulation, and using the predicted sensory results, it can be evaluated whether the desired goal is achieved. In particular, the task in this paper is to roll a sphere between the fingertips of the dexterous hand of the humanoid robot TWENDY-ONE. First, a forward model for the prediction of the touch state resulting from the in-hand manipulation is developed. As it is difficult to create such a model analytically, the model is obtained through machine learning. To get real world training data, a dataglove is used to control the robot in a master-slave way. The learned model was able to accurately predict the course of the touch state while performing successful and unsuccessful in-hand manipulations. In a second step, it is shown that this simulated sequence of sensor states can be used as input for a stability assessment model. This model can accurately predict whether a grasp is stable or whether it results in dropping the object. In a final step, a more powerful grasp stability evaluator is introduced, which works for our task regardless of the sphere diameter.",
    author = "Kohei Kojima and Takashi Sato and Alexander Schmitz and Hiroaki Arie and Hiroyasu Iwata and Shigeki Sugano",
    year = "2013",
    doi = "10.1109/IROS.2013.6696705",
    language = "English",
    isbn = "9781467363587",
    pages = "2479--2484",
    booktitle = "IEEE International Conference on Intelligent Robots and Systems",

    }

    TY - GEN

    T1 - Sensor prediction and grasp stability evaluation for in-hand manipulation

    AU - Kojima, Kohei

    AU - Sato, Takashi

    AU - Schmitz, Alexander

    AU - Arie, Hiroaki

    AU - Iwata, Hiroyasu

    AU - Sugano, Shigeki

    PY - 2013

    Y1 - 2013

    N2 - Handling objects with a single hand without dropping the object is challenging for a robot. A possible way to aid the motion planning is the prediction of the sensory results of different motions. Sequences of different movements can be performed as an offline simulation, and using the predicted sensory results, it can be evaluated whether the desired goal is achieved. In particular, the task in this paper is to roll a sphere between the fingertips of the dexterous hand of the humanoid robot TWENDY-ONE. First, a forward model for the prediction of the touch state resulting from the in-hand manipulation is developed. As it is difficult to create such a model analytically, the model is obtained through machine learning. To get real world training data, a dataglove is used to control the robot in a master-slave way. The learned model was able to accurately predict the course of the touch state while performing successful and unsuccessful in-hand manipulations. In a second step, it is shown that this simulated sequence of sensor states can be used as input for a stability assessment model. This model can accurately predict whether a grasp is stable or whether it results in dropping the object. In a final step, a more powerful grasp stability evaluator is introduced, which works for our task regardless of the sphere diameter.

    AB - Handling objects with a single hand without dropping the object is challenging for a robot. A possible way to aid the motion planning is the prediction of the sensory results of different motions. Sequences of different movements can be performed as an offline simulation, and using the predicted sensory results, it can be evaluated whether the desired goal is achieved. In particular, the task in this paper is to roll a sphere between the fingertips of the dexterous hand of the humanoid robot TWENDY-ONE. First, a forward model for the prediction of the touch state resulting from the in-hand manipulation is developed. As it is difficult to create such a model analytically, the model is obtained through machine learning. To get real world training data, a dataglove is used to control the robot in a master-slave way. The learned model was able to accurately predict the course of the touch state while performing successful and unsuccessful in-hand manipulations. In a second step, it is shown that this simulated sequence of sensor states can be used as input for a stability assessment model. This model can accurately predict whether a grasp is stable or whether it results in dropping the object. In a final step, a more powerful grasp stability evaluator is introduced, which works for our task regardless of the sphere diameter.

    UR - http://www.scopus.com/inward/record.url?scp=84893741645&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84893741645&partnerID=8YFLogxK

    U2 - 10.1109/IROS.2013.6696705

    DO - 10.1109/IROS.2013.6696705

    M3 - Conference contribution

    AN - SCOPUS:84893741645

    SN - 9781467363587

    SP - 2479

    EP - 2484

    BT - IEEE International Conference on Intelligent Robots and Systems

    ER -