An analysis to modified fuzzy PID controller with adaptive scaling factors and smith predictor

Control system based on traditional PID (TPID) algorithm is one of the most effective solutions in modern industries. In the meantime, the fuzzified PID (FPID) controller has also been researched in multiple disciplines for its better robustness and transient performance. This paper is aimed to improve a versatile procedure to determine the fuzzy I/O scaling factors automatically. First, it takes revisions on one existing hybrid auto-tuning fuzzy PID (HPID-FLC) structure and generalizes its accessibility to generic linear membership. Secondly, a function between input scaling factors and response error is expressed in different orders and applied on the fuzzy inference. A brief temperature control model is established with its transfer function tuned by Ziegler-Nichols (Z-N) frequency method and the Smith Predictor technique. The simulation results reveal that the third-order adaptive HPID-FLC whose initial scaling factors are determined directly by the TPID parameters contains the best transient characteristics in terms of the smallest percentage overshoot, fastest settle time, and an appropriate tolerance on load disturbance. The modeling and simulation processes are achieved by MATLAB/Simulink and Fuzzy toolbox.