Effective Suppression of Mover-Position- Dependent Resonances Using Combined Decoupled Persistent Oscillation Compensator and Disturbance Observer Structure

This paper presents a novel resonance suppression method for mover-position-dependent resonance. In industrial equipment and robots, increasing controller bandwidth for higher performance can induce inherent resonances. Various filter-based methods have been developed and utilized to compensate for the resonances. However, in applications, such as the belt-drive loads, where the resonance frequency can change by two to three times depending on the position of the mover, suppressing the resonance using fixed-parameter filters is not effective. This paper proposes a new resonance suppression method combining a decoupled persistent oscillation compensator (DPOC) and a disturbance observer (DOB). In the proposed method, the DPOC compensates for the mover-position-dependent resonance by attenuating the magnitude of the measured velocity signal within a specific frequency range, whereas the DOB compensates for external disturbances. Therefore, the control system with the proposed method can enhance control performance by automatically adapting to both resonance characteristics dependent on the mover position and the external disturbances. The implementation of the proposed method in the discrete-time domain is also presented, and its stability and robustness are analyzed. Experiments are performed on an industrial belt-drive servo system to demonstrate the effectiveness of the proposed method. Using the proposed method leads to an 84% reduction in overshoot and a 20% reduction in tack time compared with the results of using a well-tuned set of fixed-parameter filters, demonstrating its superiority in resonance suppression.