Learning of real robot's inverse dynamics by a forward-propagation learning rule

Hiroki Mori, Yoshihiro Ohama, Naohiro Fukumura, Yoji Uno

Research output: Contribution to journalArticle

Abstract

A forward-propagation learning rule (FPL) has been proposed for a neural network (NN) to learn an inverse model of a controlled object. A feature of FPL is that the trajectory error propagates forward in NN and appropriate values of two learning parameters are required to be set. FPL has only been simulated to several kinds of controlled objects such as a two-link arm in a horizontal plane. In this work, we applied FPL to AIBO and showed the validity of FPL on a real controlled object. At first, we tested a learning experiment of an inverse dynamic of a two-link arm in a sagittal plane with viscosity and Coulomb friction by computer simulation. In this simulation, a low-pass filter (LPF) was applied to realized trajectories because coulomb friction vibrates them. From the simulation results, we found that the learning process is stable by some adequate sets of the learning parameters although it is more sensitive to the values of the parameters owing to friction and gravity terms. Finally, we tested applying FPL to motor control of AIBO's leg. The inverse dynamics model was acquired by FPL with only about 150 learning iterations. From these results, the validity of the FPL was confirmed by the real robot control experiments.

Original languageEnglish
Pages (from-to)38-48
Number of pages11
JournalElectrical Engineering in Japan (English translation of Denki Gakkai Ronbunshi)
Volume161
Issue number4
DOIs
Publication statusPublished - 2007 Dec
Externally publishedYes

Fingerprint

Robots
Friction
Trajectories
Neural networks
Low pass filters
Dynamic models
Gravitation
Experiments
Viscosity
Computer simulation

Keywords

  • AIBO
  • Forward propagation
  • Inverse dynamics
  • Motor control
  • Multilayered neural network

ASJC Scopus subject areas

  • Energy Engineering and Power Technology
  • Electrical and Electronic Engineering

Cite this

Learning of real robot's inverse dynamics by a forward-propagation learning rule. / Mori, Hiroki; Ohama, Yoshihiro; Fukumura, Naohiro; Uno, Yoji.

In: Electrical Engineering in Japan (English translation of Denki Gakkai Ronbunshi), Vol. 161, No. 4, 12.2007, p. 38-48.

Research output: Contribution to journalArticle

@article{08d9d3fb6a8f412f8ca7b69d44161175,
title = "Learning of real robot's inverse dynamics by a forward-propagation learning rule",
abstract = "A forward-propagation learning rule (FPL) has been proposed for a neural network (NN) to learn an inverse model of a controlled object. A feature of FPL is that the trajectory error propagates forward in NN and appropriate values of two learning parameters are required to be set. FPL has only been simulated to several kinds of controlled objects such as a two-link arm in a horizontal plane. In this work, we applied FPL to AIBO and showed the validity of FPL on a real controlled object. At first, we tested a learning experiment of an inverse dynamic of a two-link arm in a sagittal plane with viscosity and Coulomb friction by computer simulation. In this simulation, a low-pass filter (LPF) was applied to realized trajectories because coulomb friction vibrates them. From the simulation results, we found that the learning process is stable by some adequate sets of the learning parameters although it is more sensitive to the values of the parameters owing to friction and gravity terms. Finally, we tested applying FPL to motor control of AIBO's leg. The inverse dynamics model was acquired by FPL with only about 150 learning iterations. From these results, the validity of the FPL was confirmed by the real robot control experiments.",
keywords = "AIBO, Forward propagation, Inverse dynamics, Motor control, Multilayered neural network",
author = "Hiroki Mori and Yoshihiro Ohama and Naohiro Fukumura and Yoji Uno",
year = "2007",
month = "12",
doi = "10.1002/eej.20456",
language = "English",
volume = "161",
pages = "38--48",
journal = "Electrical Engineering in Japan (English translation of Denki Gakkai Ronbunshi)",
issn = "0424-7760",
publisher = "John Wiley and Sons Inc.",
number = "4",

}

TY - JOUR

T1 - Learning of real robot's inverse dynamics by a forward-propagation learning rule

AU - Mori, Hiroki

AU - Ohama, Yoshihiro

AU - Fukumura, Naohiro

AU - Uno, Yoji

PY - 2007/12

Y1 - 2007/12

N2 - A forward-propagation learning rule (FPL) has been proposed for a neural network (NN) to learn an inverse model of a controlled object. A feature of FPL is that the trajectory error propagates forward in NN and appropriate values of two learning parameters are required to be set. FPL has only been simulated to several kinds of controlled objects such as a two-link arm in a horizontal plane. In this work, we applied FPL to AIBO and showed the validity of FPL on a real controlled object. At first, we tested a learning experiment of an inverse dynamic of a two-link arm in a sagittal plane with viscosity and Coulomb friction by computer simulation. In this simulation, a low-pass filter (LPF) was applied to realized trajectories because coulomb friction vibrates them. From the simulation results, we found that the learning process is stable by some adequate sets of the learning parameters although it is more sensitive to the values of the parameters owing to friction and gravity terms. Finally, we tested applying FPL to motor control of AIBO's leg. The inverse dynamics model was acquired by FPL with only about 150 learning iterations. From these results, the validity of the FPL was confirmed by the real robot control experiments.

AB - A forward-propagation learning rule (FPL) has been proposed for a neural network (NN) to learn an inverse model of a controlled object. A feature of FPL is that the trajectory error propagates forward in NN and appropriate values of two learning parameters are required to be set. FPL has only been simulated to several kinds of controlled objects such as a two-link arm in a horizontal plane. In this work, we applied FPL to AIBO and showed the validity of FPL on a real controlled object. At first, we tested a learning experiment of an inverse dynamic of a two-link arm in a sagittal plane with viscosity and Coulomb friction by computer simulation. In this simulation, a low-pass filter (LPF) was applied to realized trajectories because coulomb friction vibrates them. From the simulation results, we found that the learning process is stable by some adequate sets of the learning parameters although it is more sensitive to the values of the parameters owing to friction and gravity terms. Finally, we tested applying FPL to motor control of AIBO's leg. The inverse dynamics model was acquired by FPL with only about 150 learning iterations. From these results, the validity of the FPL was confirmed by the real robot control experiments.

KW - AIBO

KW - Forward propagation

KW - Inverse dynamics

KW - Motor control

KW - Multilayered neural network

UR - http://www.scopus.com/inward/record.url?scp=34648827523&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34648827523&partnerID=8YFLogxK

U2 - 10.1002/eej.20456

DO - 10.1002/eej.20456

M3 - Article

VL - 161

SP - 38

EP - 48

JO - Electrical Engineering in Japan (English translation of Denki Gakkai Ronbunshi)

JF - Electrical Engineering in Japan (English translation of Denki Gakkai Ronbunshi)

SN - 0424-7760

IS - 4

ER -