Swing-Up Control of Double Inverted Pendulum Based on Adaptive Neuro-fuzzy Inference System

Abstract

Neuro-fuzzy controller based on Adaptive Neuro-FuzzyInference System (ANFIS) is proposed for swinging up a double pendulum from its hanging position to the inverted position. The controller is designed through two stages. In learning stage, the controller, which is an inverse mapping model from state variable to controlled quantity, is constructed off-line. While in control stage, control output is derived from the built model. To obtain training samples for inverse mapping model, hybrid genetic algorithms are applied to optimize computation of state variables and control law during the swing-up. Simulation results showed the effectiveness of the controller and the theoretical discussions.

A Comparative Study for Position Regulation and Anti-Swing Control of Highly Non-Linear Double Inverted Pendulum (DIP) System Using Different Soft Com

“…The study considered a takagi-sugeno based ANFIS controller to provide stability at unstable positions. Yao and Chai (2007) developed a two stage ANFIS controller for swinging up DIP from hanging position to inverted position. Initially an offline inverse mapping model of the system was constructed and finally a control output is derived from the built model.…”