Control of double inverted pendulum (DIP) using fuzzy hybrid adaptive neuro controller

This paper presents a new methodological approach for selection of appropriate type and number of Membership function (MF´s) for the effective control of Double Inverted Pendulum (DIP). A Matlab-Simulink model of the system is built using governing mathematical equations. The relation between error tolerance of successive approximations and the number of MF´s for controllers is also shown. Stabilization is done using Fuzzy and Adaptive Neuro Fuzzy Inference System (ANFIS) controllers having triangular and gbell MF´s respectively. The proposed ANFIS and fuzzy controller stabilizes DIP system within 2.5 and 3.0 seconds respectively. All the three controllers have shown almost zero amount of steady state error. Both the controllers gives excellent result which proves the validity of the proposed model. ANFIS controller provides better results as compared to fuzzy controller. Results for Settling time (s), Steady state error and Maximum overshoot (degrees) for each input and output are elaborated with the help of graphs and tables.