Abstract—The regular fuzzy system can not change the expert’s experience into inference rule storage and is short of effective method to improve the membership function. Adaptive neural fuzzy inference system combines fuzzy logic and neural unit and tunes the precondition parameter and conclusion parameter with backpropagation algorithm and least-square method and can produce fuzzy rules automatically. This paper gives an adaptive-fuzzy-neural controller based on this and applies it to superheated temperature control of boiler. In the simulation, we compare it with fuzzy controller and neural network controller. The result shows that: this method improves dynamic property, steady precision and anti-jamming characteristic.

Index Terms—adaptive-fuzzy-neural, superheated temperature, fuzzy controller; neural network controller

Cite: Peifeng Niu, Guoqiang Li, and Mizhe Zhang, "Design Research of an Adaptive-Fuzzy-Neural Controller," Journal of Advances in Information Technology, Vol. 2, No. 2, pp. 122-127, May, 2011.doi:10.4304/jait.2.2.122-127