The low stiffness has significantly limited the applications of robots in high-precision processes, and existing robot stiffness identification usually requires expensive equipment or tedious manual procedures. This paper aims to present an easy-hand and cost-effective stiffness identification method to improve the positioning accuracy of robots.
This work proposes a stiffness identification method for robots with a multiple-point contact measuring instrument, which utilizes six displacement sensors to measure the pose of the end-effector with high precision. The stiffness is then identified from the kinetoelastic model of the robot with the virtual joint method (VJM), which captures the coupling between kinematics and structural flexibility.
An experimental study utilizing a UR10 robot has been conducted to verify that the position accuracy of the robot improved from 0.34 mm to 0.11 mm by considering the identified joint deflections. At the same time, a verification experiment using a Coordinate Measuring Machine (CMM) demonstrated that the distance error of the robot can be reduced from over 0.35 mm to about 0.28 mm. The result indicates that the proposed method is successfully performed on the robot.
Benefited from the proposed instrument, the method discussed in this paper has shown to be cost-effective and is capable of being implemented without time-consuming manual intervention. This method also can accurately determine the positioning error of the end-effector under heavy loads.
