Designing Full-Order Sliding Mode Controller Based on ANFIS Approximator for Uncertain Nonlinear Chaotic Systems

In this paper, a full-order sliding mode controller, based on adaptive neuro-fuzzy inference system is proposed as approximator, for controlling nonlinear chaotic systems in presence of uncertainty. At first, the full-order sliding mode controller is designed for the system in the absence of uncertainty such that the system states are converged to the sliding surface. Then, adding uncertainty to system equations, convergence of the method is illustrated using simulations. By assuming that a part of the system dynamics is uncertain and only input-output data is partly available, adaptive neuro-fuzzy inference system is used in off-line mode to approximate the uncertain dynamics of the system based on input-output data. The proposed method can effectively solve the problems of the sliding-based methods, such as chattering phenomenon and singularity. The simulation results, applied to the well-known nonlinear systems namely PMSM and plasma torch systems when they behave in chaotic mode, demonstrate effectiveness and fidelity of the proposed control method.