Grasping Control for Kinematically Redundant Parallel Robots With a Remotely Operated Gripper

摘要

This article proposes a novel grasping control method for a kinematically redundant parallel (KRP) robot with a remotely operated gripper. The proposed method consists of a motion controller and a force controller acting on each leg of the robot. The motion control model applies the computed torque method, and the force control model is based on the forces applied by each leg to the platform. These forces are decomposed into two main components: The first component generates the grasping forces, while the second one balances the platform. Since the different force components are calculated independently, the grasping forces are not coupled with others. As a result, the grasping forces can be precisely controlled without using any force/torque sensor. The grasping process introduces actuation redundancy and changes the topology of the KRP robot. In such a situation, there are infinitely many solutions for the forces in the robot, and three approaches are proposed to resolve the overconstraint. Once the forces applied on each leg are obtained, the actuated torques can be calculated based on the static force model of the leg. Finally, experiments are conducted on a prototype to verify the performance of the proposed control method.