Dual-Arm In-Hand Manipulation with Parallel Grippers Using Tactile Feedback

Pick and place tasks of large and heavy objects are challenging for robots and typically require a multi-robot system. On one hand, the availability of multiple robots can increase dexterity, on the other hand, this dexterity might be strongly limited if only tight grasps are allowed. Grasping objects with unilateral constraints using simple parallel grippers is quite demanding from a control point of view. This paper relies on tactile feedback to perform the control of the grasp forces in such a way both tight grasp and controlled sliding are possible. This in-hand manipulation ability allows the robotic system to perform both translational and rotational motions of the objects in pick and place tasks with a simple inverse kinematics approach. This manipulation dexterity avoids re-grasping actions, which are often not compatible with task or environment constraints.