An Artificial Intelligence Based Synthesis of Electrohydraulic Servo

The purpose of this work is to design a neurofuzzy controller, which can be used in a microcomputer to control the position of an electrohydraulic servo actuating primary flight controls. More specifically, an artificial intelligence based neuro-fuzzy controller is synthesized to solve a robust servomechanism problem. Indeed, the artificial intelligence based synthesis of control law provides a suitable tool to evade the mathematical model of the system, in which originates the most troubles relating to the robustness of the system. In the servomechanism problem, the position of the loaded working piston is required to minimize the tracking error. Herein, the objective function supposes a trade-off between the tracking error and the piston velocity. The second component of the objective function can be penalized from an energy reducing perspective. It was obtained a neuro-fyzzy control with a switching structure, the role of fuzzy control being supervisory antisaturating and antichattering one. Although the neurofuzzy controller design does not require of a electrohydraulic servo model, a nonlinear one was considered in numerical simulation. A discussion is finally made concerning the comparison with a similar method developed by solving of a Davison type linear robust servomechanism problem.