Control strategies for rodotic contact tasks: An experimental study

In this article, we study the contact instability problem encountered in robotic manipulators while trying to make contact with an environment, such as grasping or pushing against objects, and propose a unified control strategy capable of achieving a stable contact against both stiff and compliant environments. The problem has three distinct stages of the contact task. In the first stage, free-space motion, the robot is approaching the environment; in the second stage, post-contact force regulation; in the third, impact stage, the transition from the first stage to the second. We make an experimental comparison of the control schemes that may be used for the three stages. For example, during impact, the manipulator should not lose contact with the environment, nor exert high impulsive forces on the environment, and in the post-impact phase, the robot should have a fast force trajectory tracking. The best strategies for the above stages are experimentally determined and then combined into a single unified controller that can achieve stable contact as well as a fast force trajectory tracking response for surfaces of variable stiffnesses. This control scheme does not require a priori knowledge of the stiffness of the environment, and is able to estimate the environmental stiffness and tune gains accordingly so as to achieve the best response. Also experimentally compared is the use of such a scheme with impedance control, another method proposed in the literature for robotic contact task control. © 1995 John Wiley & Sons, Inc.