ANFIS controller for double inverted pendulum
Sheng Qiang, Qing Zhou, Xin Gao, Shuanghe Yu · 2008
A mathematical model for the double inverted pendulum is first presented. Then adaptive neural fuzzy inference system based on the state variables fusion is proposed for the control of the double inverted pendulum. The controller uses error back-propagation and least square estimator hybrid training algorithm to adjust the membership functions of each variable, optimize fuzzy rules, and identify the double inverted pendulum. Results show that proposed ANFIS can make the inverted pendulum stable.