Unified Study of Model-Informed Extensions of ADRC and State-Feedback Control with Integral Compensator

Abstract

This paper investigates generalized model-informed extensions of active disturbance rejection control (ADRC) and disturbancecentric state-feedback control with integral compensator (IDRC). Four control structures are considered, including literature-established matrixbased generalized ADRC, previously proposed regression-based generalized ADRC, and their IDRC counterparts employing corresponding embedding mechanisms. Closed-loop transfer-function representations and parametric Routh–Hurwitz stability analyses are derived forsecondand third-order integrating systems with a single inertia term under model uncertainty and observer bandwidth variations. The obtained results indicate that the embedding mechanism has a significant influence on closed-loop robustness. In particular, regression-based formulations improve stability properties for relatively slow observers but may become more sensitive to modeling errors and phase lag for large observer bandwidths. The consistency of these trends is additionally confirmed for a third-order Ball Balancing Table system including measurement filtering. Experimental results obtained on a real laboratory setup remain consistent with the predicted stability regions and confirm similar robustness relationships for both ADRC- and IDRC-based structures.