Equilibrium point control of a robot manipulator using biologically-inspired redundant actuation system

Abstract

Unlike their robotic counterparts, humans excel at various contact tasks even in unknown environments by utilizing their ability to adaptively modulate the arm impedance. As one of many theories in human motor control, the equilibrium point control hypothesis suggests that multi-joint limb movements can be achieved by shifting the equilibrium positions defined by the central nervous system and utilizing the spring-like property of the peripheral neuromuscular system. To generate human arm-like compliant motion, this study implements the equilibrium point control on a robot manipulator using redundant actuation: two actuators are installed on each joint: one to control the joint position and the other to control the joint stiffness, respectively. With the double-actuator unit, the equilibrium position and stiffness (or impedance) can be independently programmed. Also, it is possible to estimate the contact force based on angle measurement with a user-specified stiffness. These features enable the robot manipulator to execute stable and safe movement in contact tasks. A two-link manipulator equipped with the double-actuator units was developed, and experimental results from teleoperated contact tasks show the potential of the proposed approach.