Inverse Kinematics Algorithm for Dexterous Manipulation

A well-known inverse kinematics algorithm for robot manipulator has been re-adapted here and here to perform as both a planner and a kinematic controller for dual- arm/hand dexterous manipulation tasks. This method allows assigning the object motion directly and implicitly retrieve the motion of the single fingers. Moreover, in the proposed framework, it is possible to exploit the redundancy of the whole system in contact with the object, to ensure dexterity and perform stable grasps. Such a method has been employed as a planner in here, where a suitable parallel/force control has been tested in the GraspIt! environment to prove its performance. A theoretical proof shows how such parallel controller can also deal with non-planar surfaces.