A Case Study of Gravity Compensation for da Vinci Robotic Manipulator: A Practical Perspective
Other conference paper


摘要An improved method for gravity compensation of da Vinci Robot manipulator is proposed in this paper. The existing method provided by [1] is a single-step least square estimation (LSE) approach. This method can operate precisely around the center of the workspace, however, for certain configurations, the predicted errors can increase significantly and the estimation can sometimes fail to converge to desired positions. We believe these configuration dependent performance are mainly due to the following two reasons: (1) the use of a Simple Manipulator Model (SMM) without considering the parallelogram mechanism; (2) uneven mass distributions across the manipulator. To improve the performance, we propose to use an improved Manipulator Model with Parallelogram Mechanism (MMPM) for the parameter estimation. During the estimation, an iterative least square estimation (LSE) method is presented to improve the parameter estimation for the distal joints/links. Followed the initial parameter estimation, we also perform additional calibration to fine tune the output torques to compensate for model uncertainties in the physical hardware. We hope this paper can serve as a case study on how to apply gravity compensation effectively.
著者K. Lu, H. B. Lin, C. W. Vincent Hui, K. W. Samuel Au
會議名稱IEEE ICRA 2018 Workshop on "Supervised Autonomy in Surgical Robotics"

上次更新時間 2020-08-05 於 15:56