AI-based discrete sliding-mode control for real-time implementation

Abstract. Combination of sliding-mode control and an AI block based on fuzzy logic is introduced.This hybrid control structure provides very good control performance and robustness of control loops.Based on the Lyapunov™s direct method, the sliding mode control is designed to guarantee the stabilityand good tracking performance of controlled systems. The fuzzy logic block is designed to mimic thebehavior of a vehicle driver who attempts to maintain a constant speed while travelling on a roadwith ascending and descending sections. The fuzzy logic block acts as a dumper on the systemtransients and enhances the robustness of the control loop. The control algorithm is developed forimplementation in demanding real-time control applications. The new controller is compared viasimulation experiments to the optimally tuned PI controller and the pure sliding mode controller. Itshows the best performance of all three compared controllers. These Þndings were corroborated in thereal-time speed control experiments with the induction motor test-bed system.