Nonlinear robust control for robot compliant manipulation on dynamic environments

摘要

This paper investigates how to control a robot manipulator, which has model uncertainties, to perform contact tasks upon an uncertain dynamic environment. In order to perform the contact tasks successfully, the robot control system has to adjust its frequency bandwidth with respect to the dynamics of its environment. In this paper, we will present a nonlinear robust model matching control approach. We first consider the environment uncertainties and propose a nonlinear method to adjust the frequency bandwidth of the reference model as well as the robot's position controller based on the information of the interaction force, such that the robot maintains contact with its real environment. We then consider the robot's model uncertainties and apply the nonlinear robust control theory, such that the robot control system matches the reference model, robustly. This approach measures the interaction force to adjust the robot's position controller and eliminates the construction of a complex force feedback compensator. Experiments of a robot, with 1 d.o.f., pushing a dynamic object have been performed to show the effectiveness of this approach.