Automatic Grasping Position Adjustment for Robotic Hand by Estimating Center of Gravity Using Disturbance Observer

Recently, the demand for collaborative robot is increasing. Most of them are smaller to conventional industrial robots. In order to support humans in various field, it is required that they acquire flexible and dexterous motions. The soft motion based on force control will be one of keys to realize such a motion. However, the introduction of force control leads to a decrease in the rigidity of the system, which results in a decrease in positioning accuracy. This poor position tracking accuracy could cause the drop of grasped object, and it leads property damage accidents. Recently, there have been much discussions about the responsibility, handling, and risk management of accident caused by robots. As a solution, we propose the automatic grasping position adjustment for robotic hand to prevent the grasped object from falling. By observing the center of gravity and adjusting its position, the system attempts to shift unstable grasping position to stable position. This paper shows the automatic grasping position adjustment method. Then the performance of proposed method is verified by using actual robotic hand.