Fuzzy rules emulated network and its application on nonlinear control systems

In this paper, the discrete-time nonlinear systems identification and control based on an adaptive filter are introduced. This adaptive filter is implemented using the adaptive network called Multi Input Fuzzy Rules Emulated Network (MiFren). Inspired by the neuro-fuzzy network, the structure of MiFren resembles the human knowledge in the form of fuzzy If-Then rules. The initial value of MiFren's parameters can be easily selected based on the human knowledge. Then the on-line adaptive process is performed to fine tune these parameters, the convergence of the adaptive process is proven by using Lyapunov-theory-based Adaptive Filtering (LAF). In the control system application, MiFren is applied to control various selected nonlinear systems together with the proposed control law. Computer simulation results indicate that the proposed controller is able to control the target systems satisfactory.