Abstract

The industrial robots are widely employed in various industries. Normally, the industrial robots are highly repeatable. However, their accuracy is relatively poor. If the robot’s end-effector is required to move to a pre-calculated pose (as in off-line programming), its position error may reach a couple of millimeters. To meet the growing demand for high absolute accuracy in robotic applications such as deburring, polishing, drilling, and fastening, a lot of research work has been carried out. Robot calibration is normally used to enhance the accuracy by using external pose measurement sensors such as laser tracker. However, the
calibration procedure is long and the cost is high. Also, the accuracy enhancement is limited and the best reported accuracy is around ±0.1mm. In addition, the changing operating environment and the wear and tear of the robot affect the accuracy.
This research aims at developing a novel and cost-effective dynamic pose correction (DPC) strategy to address the above-mentioned issues on the accuracy enhancement. This strategy uses the vision system, i.e. C-Track from Creaform Inc. to measure the pose and integrates with the robotic controller, also known as visual servoing. To realize this strategy, three main research activities have been conducted.
First, the pose of the robot is obtained from the binocular sensor. The triangulation method for estimating the pose of an object is elaborated and the C-Track as a binocular sensor is introduced. In order to remove the noise from the C-Track’s measurements, the
analysis on the measured data is carried out. A root mean square (RMS) method is used to achieve reliable pose estimation.
Next, the DPC strategy is designed and simulated for an industrial robot, Fanuc M20-iA. This strategy can correct the position and orientation of the robot end-effector by using position-based visual servoing. A proportional-integral-derivative (PID) controller is proposed to achieve the dynamic pose correction. The algorithm does not need the kinematic and dynamic model of the robot. The controller is fully tested in Matlab/Simulink with robotic toolbox where Fanuc M20-iA is simulated. The simulation results validated the effectiveness of the proposed DPC.
The final research work is dedicated to the experimental testing of the proposed DPC on Fanuc M20-iA. The pose estimated from the C-Track serves as the feedback and the output of the DPC is given to the robot controller through Fanuc dynamic path modification
(DPM) function. As a result, the robot is guided to the desired pose in real time. The experimental results demonstrate that the robot can achieve the position accuracy ±0.05mm and orientation accuracy ±0.05 degree.