李英立, 侯澈, 罗阳, 赵忆文, 赵新刚. 一种基于谐振与抗谐振特性的柔性关节参数辨识方法[J]. 机器人, 2021, 43(3): 279-288. DOI: 10.13973/j.cnki.robot.200536
引用本文: 李英立, 侯澈, 罗阳, 赵忆文, 赵新刚. 一种基于谐振与抗谐振特性的柔性关节参数辨识方法[J]. 机器人, 2021, 43(3): 279-288. DOI: 10.13973/j.cnki.robot.200536
LI Yingli, HOU Che, LUO Yang, ZHAO Yiwen, ZHAO Xingang. A Parameters Identification Method for Flexible Joints Based on Resonance and Anti-resonance Frequency Characteristics[J]. ROBOT, 2021, 43(3): 279-288. DOI: 10.13973/j.cnki.robot.200536
Citation: LI Yingli, HOU Che, LUO Yang, ZHAO Yiwen, ZHAO Xingang. A Parameters Identification Method for Flexible Joints Based on Resonance and Anti-resonance Frequency Characteristics[J]. ROBOT, 2021, 43(3): 279-288. DOI: 10.13973/j.cnki.robot.200536

一种基于谐振与抗谐振特性的柔性关节参数辨识方法

A Parameters Identification Method for Flexible Joints Based on Resonance and Anti-resonance Frequency Characteristics

  • 摘要: 为了获取柔性关节精确的物理参数,提出了一种基于系统谐振与抗谐振特性的参数辨识方法.首先建立柔性关节的数学模型,利用该模型推导柔性关节的谐振、抗谐振频率特性与待辨识参数的数学关系,然后基于此关系建立误差回归模型,设计实验采集不同负载条件下的输入与输出数据,计算得到系统的谐振、抗谐振频率及幅值,代入回归模型并基于最小二乘法求解参数.最后,通过仿真与实验将本文方法与一般的频域特性拟合方法进行对比,结果表明在含有噪声的情况下本文方法将参数辨识平均精度从75.34%提高到90.35%,方差从25.34%降低到8.07%,验证了所提方法的可行性与有效性.

     

    Abstract: In order to obtain the precise physical parameters of the flexible joint, a parameter identification method based on system resonance and anti-resonance characteristics is proposed. Firstly, the mathematical model of the flexible joint is established to derive the mathematical relationship between the resonance and anti-resonance frequency characteristics of the flexible joint and the parameters to be identified. Based on this relationship, an error regression model is established. The input/output data are collected under different load conditions through experiments, the resonance and anti-resonance frequencies of the system and their corresponding amplitudes are calculated and substituted into the regression model, and the parameters are solved by the least squares (LS) method. Finally, the proposed method is compared with the general method of fitting with frequency domain characteristics both in simulation and experiments. The results show that the average accuracy of parameter identification is improved from 75.34% to 90.35%, and the variance is decreased from 25.34% to 8.07% in the case of noises by the proposed method, which verify the feasibility and effectiveness of the proposed method.

     

/

返回文章
返回