On-Site Implementation of External Wrench Measurement via Non-Linear Optimization in 6-Axis FT Sensor Calibration and Crosstalk Compensation

This study introduces a novel calibration method for accurate external wrench measurement using a 6-axis FT sensor, designed to mimic human-like capabilities in robots. While current approaches for FT sensor calibration often rely on fixed parameters, our method models and calibrates essential parameters: bias, crosstalk, CoM, and inclination, to ensure reliable force measurements in mobile and inclined environments. A mobile manipulator installed with an FT sensor and a gripper is used to demonstrate calibration effectiveness across varying postures and inclined conditions, with non-linear optimization applied to minimize sensor-data errors. The proposed method addresses typical calibration challenges, including the effects of the end tool and base inclined, which are not commonly covered in existing methods. Results show that, on a non-inclined base, crosstalk and CoM calibration reduce average error by approximately 14% and variance by 38%. On an inclined base, our full calibration process reduces mean error by 42% and variance by 65%. These findings highlight the importance of inclined calibration for achieving accurate external force estimations, especially for mobile manipulator applications where the environment changes often.