Physical Human-Robot Interaction (pHRI) in 6 DOF With Asymmetric Cooperation

Human-robot interaction is a growing area of research as robotic applications expand into unstructured environments. However, much of the current research has focused on tasks involving limited degrees of freedom (DOF), while not allowing the human the ability to choose the DOF on which they wish to focus. In this paper, a controller that allows human-robot cooperation in six-DOF Cartesian space is presented, which allows the human to direct their focus as they desire. The developed scheme was tested using a virtual reality system while maintaining physical interaction with the robot. Overall, the subjects were 100% successful in completion of the tasks and were able to exchange leader/follower roles with the robot bidirectionally. In addition, a reinforcement learning algorithm was shown to decrease the estimated mechanical power applied by the human to exchange roles. The latter proves the efficiency of the proposed scheme and makes it a strong candidate for applications that involve sophisticated human-robot interaction in collaborative tasks found in a plethora of cases, e.g., industry, manufacturing, semi-autonomous driving, and so on.