Finite-time convergent continuous control design based on sliding mode algorithms with application to a hydraulic drive

Sliding modes impose strong robustness toward parametric plant uncertainties and disturbances and provide for accurate tracking performance in control systems. However, in physical systems the application of sliding modes may give rise to undesirable chattering of the control signal due to actuator dynamics, possible excitation of unmodelled dynamics and structural resonant modes of load systems, etc. This may be avoided by application of smoothing functions imposing boundary layers on the control constraint, or by carrying out the design in relation to the control derivative. However, such boundary layers introduce additional design parameters and actuator dynamics may not allow the desired control accuracy to be reached. In this paper continuous controllers are proposed, with the designs taking their offset in some well-known sliding controllers. The proposed controllers preserve the finite-time convergence properties known from sliding control while at the same time avoiding control chattering, however, on the cost of robustness. Experimental results confirm the announced properties when applied to a hydraulic valve-cylinder drive, and demonstrates superior performance over conventional linear controllers.