SEARCH
MENU

History of Robotics Research and Development of Japan1988Integration, Intelligence, etc.Realization of high-speed and high-precision robot motion using iterative learning control

Sadao KawamuraOsaka University
Fumio MiyazakiOsaka University
Suguru ArimotoOsaka University
To achievehigh-speed and high-precision motion of robots, feedforward control inputis important. For the computed torquemethod based on parameterestimationof the robotdynamics, parameter estimation errors are generated. In iterative learning control (ILC), precise parameter estimation of the robot dynamics is unnecessary. The feedforward input torque pattern of ILC is formed by practical motion trials. The feedforward input torque pattern is modified solely by the motion error of the previous motion trial. Forrigidlinkrobotsystems, it has been proven that the input pattern should be modified by acceleration errors (1) or velocity errors (2). Because this result is related to the passivity of systems, ILC is interesting as a control method for use in nonlineardynamics. Furthermore, an ILC method for both position and force control [4] has been proposed along with a control method based on a combination of time axis transformation and ILC [6]. Details of the theoretical frameworkof the convergenceto the desired motion pattern, ILC withgeometricconstraintconditions, theuseof a forgetting factor, and other topics are described [7].

Movies

photo
Iterative Learning Control (ILC)

Correspondence papers

Sadao Kawamura, Fumio Miyazaki and Suguru Arimoto：Proposal of Betterment Process: A Learning Control Method for Dynamical Systems

Transactions of the Society of Instrument and Control Engineers, Vol.22, No.1, pp. 56-62, (1986) (in Japanese)
(リンク先：J-Stage Transactions of the Society of Instrument and Control Engineers (https://www.jstage.jst.go.jp/article/sicetr1965/22/1/22_1_56/_pdf)）

Sadao Kawamura, Fumio Miyazaki and Suguru Arimoto：Learning Control of Motion for Robot Manipulator

Transactions of the Society of Instrument and Control Engineers, Vol.22, No. 4, pp. 443-450, (1986) (in Japanese), (1986)
(リンク先：J-Stage Transactions of the Society of Instrument and Control Engineers (https://www.jstage.jst.go.jp/article/sicetr1965/22/4/22_4_443/_pdf)）

Related papers

[1] Suguru Arimoto, Sadao Kawamura and Fumio Miyazaki,"Bettering　Operation　of Robots by Learning", Journal of Robotic Systems, Vol.1, No. 2, pp. 123-140, (1984)

[2] Sadao Kawamura, Fumio Miyazaki, and Suguru Arimoto, "Iterative　Learning Control for Robotic Systems", Proc. of IECON'84, Tokyo, pp. 393-398, (1984) (in Japanese)

[3] 川村貞夫，川村竜也，藤野大助，宮崎文夫，有本　卓, "運動パターン学習による二足歩行ロ ボットの歩行実現," Journal of the Robotics Society of Japan，Vol.3, No.3, pp. 5-15, (1985) (in Japanese)

[4] 川村貞夫，松森正史，松林成彰，宮崎文夫，有本　卓, "学習方式による位置と力のハイブリ ッド制御," Journal of the Robotics Society of Japan, Vol. 5, No. 2, pp. 25-36, (1987) (in Japanese)

[5] Sadao Kawamura, Fumio Miyazaki, and Suguru Arimoto, “Realization of Robot Motion Based on a Learning Method”, IEEE Tran. on SMC Vol.18, No.1, pp. 126-134, (1988)

[6] Norihisa FUKAO and Sadao KAWAMURA, "A Nonlinear Time Scale Translation for Input Torque Patterns of a Robot Manipulator Considering the Dynamics," Transactions of the Society of Instrument and Control Engineers, Vol. 32, No. 7, pp. 1107-1112, (1996) (in Japanese)

[7] Suguru Arimoto,　"Control Theory of Non-linear Mechanical System: Passivity-based and Circuit-theoretic Approach" Clarendon Press Oxford, (1996)

Related Article

記事はありません。