摘要
针对进行详细的动态特性与能耗分析以及基于惯性力与重力补偿的精确控制时需要获取离散惯性参数的问题,提出了一种辨识模型及方法,研究了多轴工业机器人离散惯性参数的一次性整体辨识。为了解决超定的离散参数辨识线性方程组容易近似奇异导致最小二乘解不可靠的问题,基于观测矩阵奇异值分解与有效秩,给出了一种扩展的线性方程组的最小二乘解公式。最后,对KUKA KR60-3机器人的动力学参数进行了辨识。结果表明,离散参数对应的力矩预计结果与传统最小参数对应的结果几乎一致,且与实验结果基本相符,验证了所提辨识模型和方法的可靠性。
Aiming at the problem that discrete inertia parameters need to be obtained during detailed dynamic characteristics and energy consumption analysis and precise control based on inertial force and gravity compensation,an identification model with its method was proposed.The one-time global identification of discrete inertia parameters of multi-axis industrial robots was studied.To solve the unreliable problem of the least squares solution due to the approximate singularity of the overdetermined linear equations of discrete parameter identification,an extended least squares solution formula of linear equations was given based on the singular value decomposition of the observation matrix and its effective rank.Finally,the dynamic parameter identification experiment of a KUKA KR60-3 robot was carried out.The results show that the predicted results of torques of discrete parameters are almost consistent with those of traditional minimum parameters,which are in good agreement with the experimental results.The reliability of the proposed identification model and method is verified.
作者
周进
曹华军
江沛
吴延
兰云坤
ZHOU Jin;CAO Huajun;JIANG Pei;WU Yan;LAN Yunkun(State Key Laboratory of Mechanical Transmissions(Chongqing University),Chongqing 400044,China;Chongqing Hongyu Precision Industry Group Co.,Ltd.,Chongqing 402760,China)
出处
《中国科技论文》
CAS
北大核心
2023年第2期224-230,共7页
China Sciencepaper
基金
重庆市技术创新与应用示范专项重点示范项目(cstc2018jszx-cyzdX0163)
国家自然科学基金资助项目(51705050)。
关键词
工业机器人
动力学参数辨识
拉格朗日法
奇异值分解
最小二乘法
industrial robot
dynamic parameter identification
Lagrangian method
singular value decomposition
least squares method
