Sliding mode-based learning control for complex systems with dynamic fuzzy models

Complex systems are composed of interconnected heterogeneous components. The interconnections between these components may be nonlinear and unknown. In order to achieve precise control and control over wide operating ranges in the complex system, this thesis is concerned with the control design strategies via Takagi-Sugeno (T-S) fuzzy models. In this thesis, sliding mode learning control has been developed to address the issues of fuzzy dynamic modelling and various types of robust controller designs. The proposed learning control algorithms have three major advantages: (i) the information of the parameter variations and disturbances is no longer required in the proposed learning controller design, (ii) the control input is chattering-free, and (iii) the sliding mode learning control system possesses a strong robust property against parameter variations and disturbances.