An Experimental Study on Improvement of Weaving Motion of Welding Robots by Learning

Satoshi Tadokoro; Naoto Kobayashi; Nobukazu Kawasaki; Noriaki Miyazaki; Toshi Takamori

doi:10.1299/kikaic.61.4021

An Experimental Study on Improvement of Weaving Motion of Welding Robots by Learning

Satoshi Tadokoro, Naoto Kobayashi, Nobukazu Kawasaki, Noriaki Miyazaki, Toshi Takamori

INF - Applied Information Sciences

Research output: Contribution to journal › Article › peer-review

Abstract

In this paper, weaving trajectories of welding robots are improved by incorporating learning methods. Using functions for evaluating trajectory error based on data from accelerometers, a learning method of one-dimensional optimization and that based on a model of weaving amplitudes and phases were adopted. Experimental results showed that learning methods effectively improved the weaving trajectories of welding robots; accurate weaving trajectories were obtained by only 30 learning repetitions at most. It was also revealed that continuous weaving motion along welding lines was possible by interpolation of parameter values.

Original language	English
Pages (from-to)	4021-4026
Number of pages	6
Journal	transactions of the japan society of mechanical engineers series c
Volume	61
Issue number	590
DOIs	https://doi.org/10.1299/kikaic.61.4021
Publication status	Published - 1995

Keywords

Learning Control
Manipulator
Motion Control
Robot
Weaving Motion
Welding

ASJC Scopus subject areas

Mechanics of Materials
Mechanical Engineering
Industrial and Manufacturing Engineering

Access to Document

10.1299/kikaic.61.4021

Fingerprint Dive into the research topics of 'An Experimental Study on Improvement of Weaving Motion of Welding Robots by Learning'. Together they form a unique fingerprint.

Cite this

@article{0867bdbda4264f7d82f2001b04bc5e82,

title = "An Experimental Study on Improvement of Weaving Motion of Welding Robots by Learning",

abstract = "In this paper, weaving trajectories of welding robots are improved by incorporating learning methods. Using functions for evaluating trajectory error based on data from accelerometers, a learning method of one-dimensional optimization and that based on a model of weaving amplitudes and phases were adopted. Experimental results showed that learning methods effectively improved the weaving trajectories of welding robots; accurate weaving trajectories were obtained by only 30 learning repetitions at most. It was also revealed that continuous weaving motion along welding lines was possible by interpolation of parameter values.",

keywords = "Learning Control, Manipulator, Motion Control, Robot, Weaving Motion, Welding",

author = "Satoshi Tadokoro and Naoto Kobayashi and Nobukazu Kawasaki and Noriaki Miyazaki and Toshi Takamori",

year = "1995",

doi = "10.1299/kikaic.61.4021",

language = "English",

volume = "61",

pages = "4021--4026",

journal = "Nihon Kikai Gakkai Ronbunshu, C Hen/Transactions of the Japan Society of Mechanical Engineers, Part C",

issn = "0387-5024",

publisher = "Japan Society of Mechanical Engineers",

number = "590",

}

TY - JOUR

T1 - An Experimental Study on Improvement of Weaving Motion of Welding Robots by Learning

AU - Tadokoro, Satoshi

AU - Kobayashi, Naoto

AU - Kawasaki, Nobukazu

AU - Miyazaki, Noriaki

AU - Takamori, Toshi

PY - 1995

Y1 - 1995

N2 - In this paper, weaving trajectories of welding robots are improved by incorporating learning methods. Using functions for evaluating trajectory error based on data from accelerometers, a learning method of one-dimensional optimization and that based on a model of weaving amplitudes and phases were adopted. Experimental results showed that learning methods effectively improved the weaving trajectories of welding robots; accurate weaving trajectories were obtained by only 30 learning repetitions at most. It was also revealed that continuous weaving motion along welding lines was possible by interpolation of parameter values.

AB - In this paper, weaving trajectories of welding robots are improved by incorporating learning methods. Using functions for evaluating trajectory error based on data from accelerometers, a learning method of one-dimensional optimization and that based on a model of weaving amplitudes and phases were adopted. Experimental results showed that learning methods effectively improved the weaving trajectories of welding robots; accurate weaving trajectories were obtained by only 30 learning repetitions at most. It was also revealed that continuous weaving motion along welding lines was possible by interpolation of parameter values.

KW - Learning Control

KW - Manipulator

KW - Motion Control

KW - Robot

KW - Weaving Motion

KW - Welding

UR - http://www.scopus.com/inward/record.url?scp=0029389536&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0029389536&partnerID=8YFLogxK

U2 - 10.1299/kikaic.61.4021

DO - 10.1299/kikaic.61.4021

M3 - Article

AN - SCOPUS:0029389536

VL - 61

SP - 4021

EP - 4026

JO - Nihon Kikai Gakkai Ronbunshu, C Hen/Transactions of the Japan Society of Mechanical Engineers, Part C

JF - Nihon Kikai Gakkai Ronbunshu, C Hen/Transactions of the Japan Society of Mechanical Engineers, Part C

SN - 0387-5024

IS - 590

ER -