Robustness improvement of a model reference and sensitivity model-based self-learning fuzzy logic controller

In this paper, improved robustness of a PD type self-learning fuzzy logic controller (SLFLC) is described. The SLFLC utilizes sensitivity model for learning of SLFLC parameters. The robustness has been improved by adding an integral term whose gain coefficient is also synthesized by learning that is activated after completion of the main learning algorithm. The effectiveness of the improved SLFLC has been tested by simulation on the models of static linear and nonlinear systems. The results have proved that the SLFLC provides desired closed-loop behavior and eliminates a steady-state error in the presence of external disturbance.