Abstract: In view of the poor adaptability of the classic PID control method based on the object's exact model, which is difficult to adapt to the controlled object with nonlinearity and time-varying uncertainty, a simple PID adaptive control method based on RBF neural network is proposed. The intelligent PID control is applied to the pneumatic hydraulic servo system. The experimental results show that the RBF network PID control method with self-learning and adaptive ability can adapt to the changes of the controlled object in a large range, has strong robustness, and its control quality is significantly better than the conventional PID control method. It is feasible to apply it to the pneumatic hydraulic servo system. Keywords: RBF neural network; PID control; pneumatic servo system; fuel pump regulator 1. Introduction Pneumatic systems have the advantages of low cost, energy saving, no pollution and simple structure. Therefore, pneumatic systems have been widely used in various fields [1]. However, due to the high compressibility of air, the pneumatic servo system has strong nonlinearity, which makes it difficult to obtain satisfactory dynamic response performance and steady-state accuracy using traditional linear control theory and methods. In the control of industrial processes, PID control has the advantages of intuitiveness, simple implementation and strong robustness and has been widely used. PID control is a method based on the mathematical model of the object, and is especially suitable for deterministic control systems where an accurate mathematical model can be established. However, many industrial processes often have nonlinear and time-varying uncertainties, making it difficult to establish an accurate mathematical model. In addition, the parameters in conventional PID controllers are usually manually tuned. Since the parameters obtained by tuning at one time are difficult to achieve the best control effect, the control effect and accuracy of conventional PID controllers are limited. Therefore, researchers have been seeking adaptive techniques for PID control parameters to adapt to complex working conditions and high-performance control requirements [2]-[5]. The development of neural network technology has made this possible. However, due to the slow convergence speed, large amount of computation, and easy generation of local minima of general neural networks [6], it is difficult to apply them to control systems with high performance indicators. However, radial basis function (RBF) networks have the advantages of small amount of computation, fast convergence, and no local minima [6], making it possible to apply RBF neural networks in high-performance control systems. Therefore, the RBF neural network model was combined with the conventional PID control algorithm to form an adaptive neural network PID controller. Its application in a pneumatic and hydraulic servo system shows that the control quality of the RBF neural network PID control is superior to that of conventional PID control. For details, please click to download: Research on Intelligent PID Control of Pneumatic and Hydraulic Servo Systems.