A state estimation-based model predictive control strategy is proposed to deal with the constraints and the poor robustness problems existing in the system of servo motor driven pump-controlled asymmetric cylinders. A state-space model with input and output constraints is developed according to the principle of pump-controlled cylinder system and the practical requirements. The model predictive control is introduced to avoid the nonlinearity of valve switching and to improve the performance of constrained optimal control. The Kalman filter is combined to compensate the impacts of disturbance and modelling errors and to improve the robustness of the system. Simulation results of the strategy show that offset-free reference tracking is achieved

no over-limit of the input and no overshoot of the output are guaranteed

and model mismatch and disturbance are handled satisfactorily. Experimental results with the strategy show that the position has no overshoot with a relative tracking error of about 0.5% and the effects of disturbance and modelling error are small. Hence

the proposed method is applicable to the real pump-controlled systems

improves the control performance under input and output constraints

and allows for certain disturbance and modelling error.