Robust controller design for nonlinear uncertain dynamical systems can be a challenging work. This thesis focuses on the design, analysis and implementation of a high performance fuzzy model-based adaptive robust control (FMARC) for a class of second-order nonlinear uncertain systems (e.g., the robot manipulators), in the presence of structured (parametric) and unstructured uncertainties (e.g., disturbances). In order to provide a simple yet efficient dynamic model, modelling based on the theoiy of fuzzy logic is selected as a modelling tool to construct the parameterized fuzzy model of the system. In this regard a systematic fuzzy modelling methodology is developed. Sliding mode control methodology is selected as a framework to construct the control law (i.e., formulation to compute the required control input) and address the stability and robustness of the overall closed-loop system. The proposed approach effectively combines the design techniques from sliding mode control (SMC), adaptive control (AC), fuzzy logic (FL) modelling methodology and linear proportional-derivative-integral (PID) control to improve the performance and enhance the robustness property of the controller.

Two adaptation mechanisms are provided to deal with structured uncertainties and unstructured uncertainties, namely:​ an adaptive fuzzy model and an adaptive performance/error-based uncertainty estimator mechanism. It is proved that for slowly time-varying uncertainties, asymptotic stability can be achieved in the presence of the two above mentioned uncertainties and without using the high gains in the feedback loop. The design is conceptually simple and uses the desired trajectoiy information rather than using expensive measured acceleration information for estimation of the uncertainty, thus it is suitable for implementation.

In the first phase of validation, the proposed FMARC algorithm is compared with Slotine and Li adaptive control and sliding mode control with boundary layer by simulation study of a two degrees of freedom (DOF) direct drive serial robot manipulator. Comparative simulation results revealed that the proposed method has superior performance in terms of tracking accuracy without increasing the control effort. In the second phase of validation, the proposed method is applied to the modelling and trajectoiy tracking control of 4-DOF wire-actuated robot manipulator. The proposed FMARC is found to provide a very good tracking performance without increasing the control effort.

The simulation and experimental results show that the proposed method overcomes the drawbacks of controllers designed based on only AC, SMC and high-gain PID control methods and its simplicity makes the approach attractive for real-time applications.