Robust control of passive-jointed robot and experimental validation using sliding mode

摘要

A robust control strategy for a passive-jointed robot using sliding mode control is investigated. Instead of an actuator, the passive joint has a brake to reduce the weight and energy consumption of the robot. While the brake is released, the passive joint is indirectly controlled by the motion of the active joint using the dynamic coupling of the robot manipulators. However, there are uncertainties associated with parameters in the links as well as in the operation of the actuator and the brake. To ensure minimum variation from the desired mode of operation in the presence of these uncertainties, a sliding mode control algorithm is developed and the performance evaluated. From simulation results, the sliding mode scheme is superior to a simple computed torque method in terms of passive-joint tracking performance and robustness. A two-link manipulator was designed and built to evaluate the proposed control scheme experimentally. The robot consists of one active joint and one passive joint and operates in a horizontal plane. The planar manipulator was controlled from a 32-bit 80486 microcomputer via DT2811 input/output board. The experimental results indicate that the sliding mode controller, when compared to the computed torque controller, reduces the tracking error by 50% and improves the robustness to variations in payloads.