High-Speed Adaptive Control of Robot Manipulator

Abstract

One of the biggest problems of realizing the fast motion of a robot manipulator is its dynamic reflection. The driving torque calculated from the mathematical model well compensates for this reflection. However, the dynamic property changes according to the clamped load and arm position. This paper introduces a high-speed adaptive control technique using identification. First, the manipulator is controlled only by a planned trajectory. A time domain identification is then applied to estimate the robot dynamics. The inverse transfer function is used to calculate the actuating signal which produces the high-speed motion of the next cycle. The proposed technique is applied to 3-dimensional positioning of a multijoint manipulator.