Robot self-calibration using multiple kinematic chains

Proper calibration is key for the performance of every robot. Pushed by societal needs and economic opportunities, robots are leaving fixed factory floors and are deployed in more versatile ways both in industry and outside, which increases the need for automated calibration procedures. At the same time, advances in sensor technology make affordable but increasingly accurate devices such as RGB-D and tactile sensors available, making it possible to perform automated self-contained calibration relying on redundant information in these sensory streams. In this work, we take the example of a humanoid robot with a stereo camera system and force-sensitive end-effectors and quantitatively compare the performance of kinematic calibration by employing different combinations of intersecting kinematic chains - either through self-observation or self-touch.