In this paper, we propose a design scheme for the intelligent control for a pneumatic servo system. In this design scheme, the unknown parameters are identified first and then the plant model is determined whether to be minimum-phase or not, because the discrete-time plant model of the pneumatic servo system becomes a minimum phase or a non-minimum phase depending on the difference in the sampling period and the load mass. Next a model reference adaptive control system (MRACS) is adopted in the case of minimum-phase and an adaptive pole-assignment control system (APACS) is used in the other case. Furthermore, in this design scheme, neural networks (NN) in the adaptive controller and a deference process of the signals in the identification mechanism are introduced to eliminate the influence of non-linear elements caused by compressibility of the air and various frictions. The effectiveness of the proposed design scheme is confirmed by experiments using an existing pneumatic servo system.
