The Neural Network Position-Control of a Moving Platform Using Electrorheological Valves

Abstract

This paper presents position control of a moving platform using a hydraulic bridge circuit associated with electrorheological (ER) valves. Four cylindrical ER valves are designed and manufactured on the basis of electric field-dependent yield stress of an ER field, which is composed of chemically treated starch and silicone oil. The pressure drops of the ER valves are empirically identified with respect to the intensity of the electric field, and the hydraulic bridge circuit with four ER valves is constructed. The hydraulic cylinder system to be controlled by the ER valve bridge circuit is then incorporated with a moving platform of a cargo handling laboratory model. Subsequently, a neural network control scheme is formulated in order to control the position of the moving platform by activating ER valves of the cylinder system. The controller is experimentally realized and position tracking control results for desired trajectories of the moving platform are presented in the time domain.