Estimation and control of non-linear plant in ARX form using MDPP controlled Fuzzy Identifier

The paper presents the estimation of an unknown non-linear system in ARX form. The estimated model is used in Adaptive self-tuning regulator to regulate the output of plant at desired reference input using Minimum Degree Pole Placement Controller (MDPP). Most of the literature includes the conventional RLS algorithm to estimate the system parameters. The estimation technique addressed here is Fuzzy Logic Identifier (FLI). The proposed estimator approximates the system using a linear model at each operating point. The linear identification of model necessitates the use of linear control strategy. Therefore, the design of MDPP controller based on linear identified system is also presented. The consequents of FLI are updated using Levenberg-Marqardt (LM) algorithm. MATLAB/Simulink is used for the implementation of proposed estimation techniques. The effectiveness of the proposed estimators and control strategy is tested on a non-linear plant. The results are presented and debated together with a comparative analysis of RLS and FLI. Superiority in performance is found for FLI in estimating and controlling the non-linear using MDPP. The performance indices Mean Square Error (MSE), Mean Absolute Error (MAE) and Mean Time-Weighted Error validate the effectiveness of proposed FLI.