In this work we show how to use the Operational Space Control framework (OSC) under joint and cartesian constraints for reinforcement learning in cartesian space. Our method is therefore able to learn fast and with adjustable degrees of freedom, while we are able to transfer policies without additional dynamics randomizations on a KUKA LBR iiwa pegin-hole task. Before learning in simulation starts, we perform a system identification for aligning the simulation environment as far as possible with the dynamics of a real robot. Adding constraints to the OSC controller allows us to learn in a safe way on the real robot or to learn a flexible, goal conditioned policy that can be easily transferred from simulation to the real robot. 1
In this paper, the problem of including hard constraints in the stack of tasks for torque-controlled serial manipulators is treated. The classic potential field approach is studied and a novel implementation of it is proposed. This implementation reduces the kinetic energy in the proximity of joint limits. Furthermore, a novel approach is proposed in order to include not only joint positions but also joint velocities and acceleration limits. This approach is called “Saturation in Joint Space” (SJS). The algorithm proceeds by creating a task with the highest priority in a stack of tasks scheme. This high priority task saturates the acceleration of the joints that would exceed their motion limits. The methods are tested and compared in simulation for the KUKA LBR iiwa. The SJS approach presents smoother behaviour near to the joint limits, while a Cartesian trajectory is traced. Experiments are performed to test the efficiency of this method in a real environment and under interaction with a human. The on-line saturation of the joint acceleration generates a friendly behaviour with the human even when he pushes the robot towards its limits.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.