Design and performance analysis of reset-based active disturbance rejection control

To address the inherent trade-off between transient performance and noise robustness of the traditional linear extended state observer (LESO) within the active disturbance rejection control (ADRC) framework, a novel reset extended state observer (RESO) scheme is proposed. Its core idea is to introduce a state correction mechanism regulated by the reset gain matrix into the LESO, which can maintain structural simplicity to the greatest extent. The reset action is activated only when the output estimation error undergoes a significant zero-crossing with the minimum reset interval constraint, thus achieving on-demand reset and avoiding Zeno behaviour. The closed-loop error dynamics model comprising flow dynamics and jump dynamics is established, and its stability is rigorously proven via the Lyapunov method. Furthermore, a systematic linear matrix inequality (LMI)-based procedure is developed for designing the reset gain matrix. Comparative simulations on a mass-spring-damper (MSD) system demonstrate that the proposed RESO achieves superior transient performance and noise robustness relative to the conventional LESO.