Neuro-fuzzy control for pneumatic servo system

A learning method for acquiring the appropriate fuzzy rules using error back propagation to improve the control performance of a pneumatic servo system is presented. In the proposed method, a criteria is defined and the fuzzy rules are adjusted so as to minimize them using error back propagation. Moreover, differentiation of the coefficient of the plant used in error back propagation is accomplished by the newly established neural network. The proposed method is applied to vertical pneumatic servo systems to prove their effectiveness