To address the challenges of vibration suppression, reference tracking, and disturbance rejection in robot servo systems, a novel model predictive control (MPC) framework based on output regulation theory and disturbance observer techniques is proposed in this paper. Firstly, the vibration mechanism of the flexible robot joint is analyzed to model the torsional torque disturbance. Then, the disturbance is estimated using a reduced-order observer. Subsequently, based on the disturbance estimation and output regulation theory, the reference trajectories for both the system state and control input are derived. Further, through the equivalent state transformation, the tracking problem of the original disturbed system is converted into a stabilization problem for the transformed system. This reformulated problem can then be solved by the proposed MPC scheme while explicitly handling state and control input constraints. The experimental results verify that the proposed MPC can effectively suppress the vibration and improve the dynamic performance of the robot servo systems.

disturbance estimation,MODEL PREDICTIVE CONTROL,output regulation,servo system,vibration suppression