FINGERTIP LOAD ESTIMATION DEVICE, ROBOT CONTROL SYSTEM AND ROBOT SYSTEM

To acquire a fingertip load estimation device capable of acquiring a highly accurate estimation result for a fingertip load parameter.SOLUTION: A fingertip load estimation device 1 includes: a reverse dynamical model part 12 for outputting a nominal torque value which is a torque calculated based on...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: MAEKAWA SEISEKI, MATSUOKA RYO, YAMANOBE NATSUKI, KAJITA HIDEJI, AKAHO SHOTARO, ASO HIDEKI
Format: Patent
Sprache:eng ; jpn
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:To acquire a fingertip load estimation device capable of acquiring a highly accurate estimation result for a fingertip load parameter.SOLUTION: A fingertip load estimation device 1 includes: a reverse dynamical model part 12 for outputting a nominal torque value which is a torque calculated based on a reverse dynamical model, by movement data which is at least one piece of data of the position, speed, acceleration of a joint of a robot being inputted in the reverse dynamical model; a learning part 13 which learns a fingertip load parameter in which the influence of a modeling error of the reverse dynamical model and the influence of the characteristic which is not included in the reverse dynamical model out of the characteristics of the robot are excluded, based on the movement data and data for learning containing a difference between an actual torque of the joint and the nominal torque; and an estimation part 15 in which data for estimation containing the movement data and the difference between the actual torque and the nominal torque is inputted, and which estimates the fingertip load parameter based on the learning result by the learning part 13.SELECTED DRAWING: Figure 1 【課題】手先負荷パラメータの高精度な推定結果を得ることができる手先負荷推定装置を得ること。【解決手段】手先負荷推定装置1は、ロボットの関節の位置、速度および加速度の少なくとも1つのデータである動きデータが逆動力学モデルに入力されることによって、逆動力学モデルを基に計算されるトルクである公称トルクの値を出力する逆動力学モデル部12と、動きデータと、関節の実トルクおよび公称トルクの差分とを含む学習用データに基づいて、逆動力学モデルのモデル化誤差の影響と、ロボットの特性のうち逆動力学モデルに含まれていない特性の影響とが除かれた手先負荷パラメータを学習する学習部13と、動きデータと、実トルクおよび公称トルクの差分とを含む推論用データが入力され、学習部13による学習結果を基に手先負荷パラメータを推論する推論部15と、を備える。【選択図】図1