Constrained incremental predictive controller design for a flexible joint robot

In this paper, an improved predictive control algorithm for controlling a typical nonlinear flexible-joint robot (FJR) with input constraint is proposed. The receding horizon algorithm, called generalized incremental predictive control (GIPC), utilizes both present and previous states rather than present states only. The GIPC algorithm includes the weighted difference of the current and the previous states and the summation of the control action increments. In order to illustrate the effectiveness of the proposed control strategy, it is implemented to the FJR and the results are compared with those of generalized predictive control (GPC). It is demonstrated that the proposed GIPC algorithm is more robust than the standard GPC method. Furthermore, the constrained GIPC algorithm using the quadratic programming removes instabilities caused by actuator saturation.