http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Machine-learning-based approach to improve the positioning accuracy of large industrial robots
K. Yoshitsugu,D. Kato,T. Hirogaki,E. Aoyama,K. Takahashi 제어로봇시스템학회 2020 제어로봇시스템학회 국제학술대회 논문집 Vol.2020 No.10
In this study, a system to reproduce the positional relationship of two flanges to be connected using an industrial robot and enable the manufacturing of the connecting pipes was developed. The objective was to ensure that the robot can exhibit a high absolute positioning accuracy even when trained offline so as to reproduce the space accurately. The error in various postures of the robot was measured using a laser tracker, and the servo information was simultaneously acquired by using the OpenNR-IF framework. The relationship between the positioning error and servo information was examined. It was noted that the relationship between multiple parameters such as the joint angle obtained as the servo information was highly complex. Therefore, the random forest approach was employed in machine learning to examine these relationships. Subsequently, the positioning error was predicted based on each parameter and corrected. Although the correction was inadequate, the mechanism to select the data to enhance the prediction accuracy was clarified, and several relevant key parameters were identified.
J.Yoshitsugu,M.Ando,M.Rukonuzzaman,E.Hiraki,M.Nakaoka,K.Inoue 전력전자학회 2001 ICPE(ISPE)논문집 Vol.2001 No.10
This paper presents a circuit of the quasi-resonant DC link to achieve soft-switching three phase inverter using intelligent IGBT power module. The soft-switching operation in this circuit is confirmed simulation and experimental results. Its conductive noise is measured for electrical AC motor drive as compared with that of the conventional hard switching inverter.<br/>