Load-side resonant suppression based on adaptive state feedback and decoupled sliding-mode observer for servo system

The engineered bi-quad notch filters successfully compensate for the flexibility in the dual-inertia servo system, effectively suppressing inherent system resonance but potentially introducing load-side oscillation. To address this, a novel load-side resonance suppression scheme employing an adaptive state feedback is proposed, enabling free adjustment of the load-side damping, which can mitigate the load-side oscillations and improve system resonance suppression efficiency. In addition, since the servo system requires accurate mechanical parameters to properly configure the state feedback, two decoupled sliding-mode observers (DSMOs) are proposed to identify load torque and load inertia simultaneously. Interestingly, each DSMO incorporates an innovative decoupling term that eliminates parameter estimation coupling with motor speed errors, significantly reducing identified value chattering. Meanwhile, an improved reaching law is also designed as the gain factor to achieve the balance between rapid identification convergence and effective chattering suppression. Comprehensive experimental validation demonstrates the effectiveness of the proposed scheme, with results confirming substantial performance improvements in resonance suppression and parameter identification accuracy.