A neuro-fuzzy method for tracking control
Ching‐Tzong Su, Guor-Rurng Lii, Hong-Rong Hwung · 2002
The purpose of this paper is to propose a new approach to be used in optimal position control. This method uses fuzzy control system and works with genetic algorithms (GAs) to meet the requirement of optimal position control. Based on the unsupervised training of self-organizing neural network, the fuzzy expert experiences are learned. The neuro-fuzzy controller (NFC) then applies, these experiences learned to determine the output control force. By virtue of the evolution rule of genetic algorithms, the best expert experiences are extracted and employed to achieve the optimal position control. Application of the proposed method to the inverted pendulum system is also presented. The simulation results show that the controller has satisfactory performance.