Performance Enhancement of Fuzzy Logic Controller Using Robust Fixed Point Transformation

Despite its excellent performance as a controller for linear and non-linear systems, the fuzzy logic controller has certain limitations. For instance, large-scale complex fuzzy systems like multi-input, single-output, or multi-output systems are used in various applications with large number of rules. Furthermore, the results also depend on the selected membership functions, etc. This paper presents a novel framework that instead of reducing the number of rules for a fuzzy logic controller, combines it with a fixed point transformation based adaptive control. The adopted approach is based on the Mamdani-type fuzzy controller and enhanced by the Sigmoid Generated Fixed Point Transformation control strategy to cope with modeling inaccuracies and external disturbances that can arise. The general procedure is applied to a nonlinear Kapitza pendulum. Numerical simulations are validating the applicability of the proposed scheme and demonstrating the controller's performance.