为了辨识具有谐波减速器的柔性关节模型的关键参数,设计了一套基于多传感器融合的离线辨识方法,根据电机的位置传感器、电流传感器和关节力矩传感器的实验数据完成柔性关节模型关键参数的辨识.首先,建立采用谐波减速器柔轮输出的柔性关节模型;然后,使用正反转加载力矩的方法辨识出电机的力矩系数;并在空载的情况下,由关节力矩和电机输出力矩分别辨识出关节端和电机端的摩擦力;最后,采用敲击法初步辨识出关节的刚度和阻尼后,在关节位置受限条件下,逐渐增加电机输出力矩,得到柔性关节刚度和关节力矩的非线性关系.多次实验的结果显示,辨识出的参数具有较高的重复性,验证了该方法的有效性.
In order to identify the key parameters of flexible joint model with harmonic driver, an off-line identification method by integrating several sensors is proposed. The key parameters can be identified by the experimental data of motor position sensors, current sensors and joint torque sensors. Firstly, the flexible joint model is established, which uses the flexspline as the output in the harmonic driver. Then, the motor torque coefficient is identified by running in forward and inverse directions with different loads. In the no-load condition, joint friction and motor friction are identified from the joint torque and motor output torque. Finally, the stiffness and damping parameters are identified by the knocking method, while the nonlinear relationship between stiffness and joint torque is obtained by progressively increasing the motor output torque under the condition that the joint position is hold. Several experiments are repeated, and the results show a high repeatability, which verifies the proposed method.