An Enhanced <i>H</i> _∞ Design of Disturbance Observer for a High-Bandwidth Fast Steering Mirror

This article proposes an enhanced disturbance observer (DOB) design method to address the high-bandwidth control challenges of flexible fast steering mirror (FSM) systems. FSMs rapidly adjust mirror orientation through precision control systems, enabling them to meet the dynamic tracking requirements of various high-precision optical applications. Due to the stringent precision demands of optical systems, FSMs are susceptible to external disturbances, sensor measurement noise, and parameter variations. DOB is widely employed in motion control systems to estimate and compensate for these disturbances. However, existing DOB design methods often rely on oversimplified models, treating dynamic characteristics such as flexible resonance as model uncertainties, which limits their performance in flexible systems. Through a comprehensive analysis, this article demonstrates that in flexible resonant systems, the disturbance estimation bandwidth of the existing DOB is frequently constrained by robust stability requirements, leading to suboptimal performance. With the increasing demands on the control bandwidth of FSMs driven by modern technological advancements, model mismatches arising from high-frequency unmodeled dynamics have become a critical bottleneck. Existing DOB is unable to effectively address model mismatches in the high-frequency range of flexible systems. This article proposes an enhanced DOB design method based on the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$H_{\infty } $ </tex-math></inline-formula> control framework and Hankel matrix system identification, which simultaneously satisfies multiple performance indicators. By precisely selecting the nominal model and estimating uncertainties, the existing DOB design method is optimized. Experimental results demonstrate that the proposed enhanced DOB effectively suppresses external disturbances and model mismatches, significantly improving the system’s disturbance rejection performance compared to existing methods.