Integrating Contact, Modeling, and Control for the Robotic Hand Manipulation

A novel manipulation control framework is proposed for the dexterous robotic hand, which integrates the constraints (contact constraints and servo constraints), dynamics modeling and controller design. In the manipulation task, the contact between the robotic hand and the target object translates the manipulating force and constraints the relative motion of the target. Hence, the contact forces can be considered as the bridge for the robotic hand manipulation and the motion control of the target object. By the Udwadia-Kalaba theory, the dynamics model of the contact forces for the rolling manipulation in the work space is constructed explicitly without any auxiliary variable and calculated by the states of the robotic hand (such as, the angular and velocity of the finger joints) and the local variables of the target object (e.g., the surface parameters). Based on the formulated contact forces, an integrated control strategy is introduced to tackle with the manipulation task by combining the desired motion of the target object and the manipulation controller design through the contact constraints. Virtue of the calculated contact forces, the proposed control could accomplish the manipulation tasks without force sensors. Besides, to ensure the trajectory of the object is smooth and continuous, a grasping plan for the dexterous robotic hand is proposed. The effectiveness of the control is verified by both theoretical proof and the numerical simulation of a three fingered robotic hand in 3D workspace.