Aspects regarding the neuro-adaptive control structure properties. Application to the nonlinear pneumatic servo system benchmark

Abstract

Generally, control techniques based on multilayer neural networks are used especially in the case of non-linear process and less in case of linear process, because classical control techniques are good enough and relatively easy to implement in linear case. In this paper it will be presented a adaptive neural control structure with application to an pneumatic servo system. This system has many advantages, such as high speed, high flexibility and low price. However, applications of this system are restricted because the physical parameters have strong nonlinearity, inaccuracy and uncertainty, so it’s very difficult to find an optimal controller by means of traditional control theory.