Fuzzy rules emulated networks with adaptive controller for nonaffine discrete-time systems

An adaptive controller for a class of nonaffine discrete-time systems is developed as the main contribution of this article. With the system’s properties obtained by the second-order Taylor expansion, muti-input fuzzy rules emulated networks or MIFRENs are implemented to approximate the unknown plant under control. The closed-loop performance is guaranteed by an on-line learning algorithm developed to tune the parameters inside MIFRENs. According to the computation management, only linear parameters are adjusted with the constraints issued by the main theorem. Furthermore, the suitable learning rate can be determined with the information provided by the MIFREN approximation. The computer simulation system demonstrates the validation of the proposed controller. Moreover, the system robustness is described both nominal system and uncertain system.