机器人结构参数辨识的两步法及其仿真

实际机器人模型同控制器中机器人模型之间的偏差影响了机器人的精度,可用标定方法校正此偏差。传统的单点标定方法,能简化标定操作但仍需测量设备,不利于现场标定。改进了单点标定方法,提出了一种现场应用中无需测量设备的单点两步标定新方法,即先标定出机器人结构参数对某一长度参数的相对值和关节码盘初始安装偏差,基于两点距离模型标定出此长度参数,得到所有的机器人实际结构参数。基于机械分析仿真平台—InteMech,设计了一套仿真试验方法,并给出了一个仿真实例。仿真试验表明该方法完全适合于机器人标定。标定模型基于D-H参数建模,通用性强,可扩展到多自由度机器人的标定。

Modeling and Simulation of Structural Parameter Identification for Robot Manipulators Using Two-step Schemes

Many calibration methods have been explored to increase the absolute accuracy of robot manipulators. Compared with others, single point contact-based calibration(SEC) decreases the calibrating operation. Nevertheless, this technique still needs measuring tools for deciding certain parameters of a manipulator. This work presents two-step calibrating schemes without measurement devices in present by improving SEC. Value of structural parameters of a robot manipulator relative to its some length parameter and initial installing deviations of joint coder was calibrated. Then, this length could be obtained based on two-point distance model. So, all parameters of the robot were determined. A simulation example was designed to validate the feasibility of this calibration method in a simulation platform of mechanism analysis-InteMech. The result shows that the way is effective and efficient for robot calibration. The great generality of the algorithm enables to be used in the calibration of multi-DOF robots.