Neural-network-based variable structure control of electrohydraulic servosystems subject to huge uncertainties without persistent excitation

• Published in: IEEE/ASME transactions on mechatronics Vol. 4; no. 1; pp. 50 - 59
• Main Author: HWANG, C.-L
• Format: Journal Article Conference Proceeding
• Language: English
• Published: New York, NY IEEE 01.03.1999; Institute of Electrical and Electronics Engineers
• Subjects: Applied sciences; Artificial intelligence; Computer science; control theory; systems; Connectionism. Neural networks; Control systems; Drives; Electric variables control; Electrohydraulic; Electrohydraulics; Exact sciences and technology; Excitation; Friction; Intelligent control; Lyapunov methods; Mathematical models; Mechanical engineering. Machine design; Mechanical variables control; Neural networks; Nonlinear control systems; Radial basis function; Servosystems; Speed variators, torque converters. Hydraulic drives and controls, pneumatic drives and controls, fluids and components, hydraulic motors, pneumatic motors; Stability; System stability; Uncertain systems; Uncertainty; Variable structure control; Vibration; Radial basis function; Time varying system; Uncertain system; Friction; Hydraulic control; Control synthesis; Neural network; Exponential stability; Electrohydraulic control
• ISSN: 1083-4435
• DOI: 10.1109/3516.752084

A novel scheme investigating a radial-basis-function neural network (RBFNN) with variable structure control (VSC) for electrohydraulic servosystems subject to huge uncertainties is presented. Although the VSC possesses some advantages (e.g., fast response, less sensitive to uncertainties, and easy implementation), the chattering control input often occurs. The reason for a chattering control input is that the switching control in the VSC is used to cope with the uncertainties. The larger the uncertainties which arise, the larger switching control occurs. In this paper, an RBFNN is employed to model the uncertainties caused by parameter variations, friction, external load, and controller. A new weight updating law using a revision of e-modification by a time varying dead zone can achieve an exponential stability without the assumption of persistent excitation for the uncertainties or radial basis function. Then, an RBFNN-based VSC is constructed such that some part of uncertainties are tackled, that the tracking performance is improved, and that the level of chattering control input is attenuated. Finally, the stability of the overall system is verified by the Lyapunov stability criterion.