The purpose of contour control is to reduce contour error, but the contour performance is limited by the system delay. Although the conventional position predictive control (C-PPC) can solve the above problem, it relies on the system parameter and ignores the coupling delay. In other words, C-PPC uses an inaccurate parameter and prediction period. To optimize the contour performance for the X-Y motion platform with parameter mismatch and system delay, a robust position predictive control (R-PPC) is proposed. Firstly, prediction error sensitivity in C-PPC is analyzed respectively when the parameter is mismatched and the prediction period is inaccurate. Then a simplified two-step predictive model is developed, which does not depend on the specific motor model and parameter. Based on this model, the robustness coefficients are introduced into the prediction model to establish R-PPC. Theoretical analysis proves that R-PPC can reduce the prediction error sensitivity by adjusting the robustness coefficients, so as to obtain more accurate position information. The simulations show that, compared with the PID controller and C-PPC, the proposed R-PPC can effectively reduce the contour error in the X-Y motion platform by improving the prediction model robustness. It means that the prediction error caused by the mismatched parameter in R-PPC is the smallest.

contour error， system delay， parameter mismatch， prediction peri-od， X-Y motion platform