A Position Control Based on Active Disturbance Rejection Controller Considering Parameter Variations of Harmonic Drive Gear Systems

The accuracy of position control is of great importance in industrial robot application. However, the control performance of traditional model-based methods is limited by the parameter variations and uncertainties. Therefore, when the harmonic drive gear undergoes load torque disturbances, the underdamped vibration will be generated in the link and a steady-state error will be existed in the system. In this case, a backstepping control structure based on active disturbance rejection controller is proposed. According to the method, additional feedback of link variables is introduced by the backstepping control to improve the system damping. In the meantime, torque disturbances, parameter variations and unmodeled dynamics are integrated and observed by the extended state observers. The observation errors can be sharply decayed and converged to zero with proper poles assigned. Combined with the feedforward compensation, the steady-state error is significantly reduced. The effectiveness of the method is validated by the simulation results.