DESIGN OF ADAPTIVE ROBUST CONTROLLERS BASED ON BACKSTEPPING TECHNIQUE FOR HYDRAULIC ROBOTIC MANIPULATORS

Abstract

This thesis focuses on studying some advanced control approaches for a hydraulic robotic manipulator with the presence of external disturbance, modeling error, and output constraints. Because hydraulic actuators are characterized by a significant power to mass ratio property, they can help to enhance the operational capability of hydraulic robots. However, the highly nonlinear dynamics concerned with hydraulic actuators challenge the control design. This thesis, therefore, presents highly effective control solutions for the selected manipulator by taking the system dynamics into account. When the actuator dynamics are considered for control development, some robust control approaches which are combined with the adaptive laws and adaptive approximators based on the backstepping. The robust approaches are used to guarantee the stability of the controlled system. In order to improve the control performance and minimize control effort, adaptive approximators are derived coping with the external disturbances and uncertainties such as modeling errors, unknown frictions, internal and external leakages in both the actuator and manipulator dynamics. Depending on types of uncertainties such as known structure, unknown structure, smooth or unsmooth uncertainties, some specified adaptive approximators are developed to handle these uncertainties. In this thesis, parametric adaptive approximators and adaptive switching laws from discontinuous control laws are respectively proposed to handle the known and unknown structure uncertainties of a 1-DOF manipulator in chapter 4. Additionally, a radial basis function neural network (RBFNN) and adaptive switching laws of robust control laws are respectively employed to deal with the smooth and unsmooth uncertainties of a 3-DOF manipulator in chapter 5. The adaptive switching laws will adjust the robust gain online based on the predefined threshold values, which help to reduce the chattering effect caused by the robust control. Furthermore, in order to improve the effectiveness of the RBFNN, adaptive laws for the weighting vectors, and parameters of RBFs are fully investigated based on the Taylor series expansion in this chapter. In both these approaches, the stability and robustness of the proposed controllers are theoretically demonstrated based on Lyapunov approaches. Additionally, simulations and experiments on a one-degree of freedom (DOF) hydraulic manipulator and three-DOF hydraulic manipulator are conducted to verify the effectiveness of the designed controllers.