Fuzzy Inference System Based on Neural Network for Technological Process Control

The implementation of a fuzzy system for technological process control based on parallel architecture and learning capabilities of neural networks is considered. The algorithms of fuzzy inference system on neural network (neuro-fuzzy system) are described. To train unknown coefficients of the system, the supervised learning algorithm is used. As a result of learning, the rules of neuro-fuzzy system are generated. The neuro-fuzzy system is applied to control a dynamic plant. Using desired time response characteristics of the system the synthesis of neuro-fuzzy controller for technological process control is carried out. The simulation result of the neuro-fuzzy control system is compared with the simulation results of control systems based on PID- and neural controller. It is found that the neuro-fuzzy control system has better control performance than the others.