Extended Asymmetric Velocity Moderation: A reactive strategy for human-safe robot control

Abstract

Human safety plays a crucial role for the symbiosis between humans and robots. Moreover, it is also important to maintain the robot's efficiency while keeping the human unharmed. With this purpose, we previously developed Asymmetric Velocity Moderation (AVM) as a reactive strategy for human safety. Nevertheless, this original AVM only considers the end-effector's movement. This may lead to overrestrictions of the robot speed when the human is very close and to underestimate human safety by assuming the end-effector is the main source of danger. Therefore, this paper extends AVM by using restrictions based on the velocities of all points on the robot and their corresponding minimum distances to the whole human body. By eliminating overrestrictions without undermining human safety and by considering the whole robot, a more efficient humansafe robot behavior can be obtained. The method proposed in this paper consists of calculating independent restrictions for every point on the robot and choosing the firmest restriction to limit the robot velocity. Simulation experiments using a virtual environment with a human model and a human-sized humanoid robot were performed for the validation of the proposed method, and its efficiency was evaluated using the task completion time of the robot.