Open AccessRealization of Swing Manipulation by 3-DOF Robot Arm for Unknown String via Parameter Estimation and Motion Generation
Author(s) -
Kenta Tabata,
Hiroaki Seki,
Tokuo Tsuji,
Tatsuhiro Hiramitsu
Publication year - 2021
Publication title -
international journal of automation technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.513
H-Index - 18
eISSN - 1883-8022
pISSN - 1881-7629
DOI - 10.20965/ijat.2021.p0774
Subject(s) - string (physics) , swing , robot , realization (probability) , computer science , control theory (sociology) , actuator , focus (optics) , damper , simulation , artificial intelligence , control engineering , engineering , mathematics , physics , control (management) , mechanical engineering , statistics , optics , mathematical physics
Dynamically manipulating flexible objects using robots is difficult. Some studies have been conducted that only considered one type of object with known properties or that needed an identification test for string properties in advance. We propose a method to realize the dynamic manipulation of a string with unknown characteristics. We use a mass-spring-damper model for the string and repeat three steps: motion generation, real manipulation, and parameter estimation. The proposed method estimates the string properties to realize the motion objective via the real manipulation of the string. An identification test in advance was not necessary. In this study, we focus on swing manipulation. This can increase the motion energy of a string without a high-power actuator. After making a large swing, the robot can throw strings to a more distant target, such as a hammer throw. This motion is useful for explanation robots, rescue robots, and so on. We modified the proposed method to generate a swing manipulation. Then, we investigate whether swing manipulation can be performed by the proposed method and demonstrate its effectiveness via experiments with various strings with unknown characteristics.