A SLAM-Integrated Kinematic Calibration Method for Industrial Manipulators with RGB-D Cameras

Abstract

The accuracy of robot manipulator, one of the long-standing problem, is a major issue in the industry community. The manipulator may produce kinematic errors during operation. Traditional methods require expensive equipment with complex steps to calibrate kinematic parameters. Another issue is motion planning of the manipulator, which requires a map of the workspace. However, the mapping is time-consuming. In order to employ an efficient way to accomplish kinematic calibration and offer convenience to plan the motions of the manipulator, we study a new method called SKCLAM (Simultaneous Kinematic Calibration, Localization, and Mapping), which can calibrate the kinematic parameters of an industrial manipulator and achieve 3D environmental mapping simultaneously by employing an RGB-D camera attached to the end effector. In this paper, the true kinematic parameters were changed randomly to test and evaluate the effectiveness of our approach in simulation.