Refining ℌ∞ Controller Performance With Extremum Seeking Control for Improved Disturbance Attenuation

Abstract

Robust disturbance attenuation is critical in automotive, active suspension systems to ensure vehicle stability and passenger comfort. Traditional H∞ control methods effectively minimize disturbances but often result in high-order controllers, which are challenging to implement practically due to complexity and reduced robustness to unmodeled dynamics. Although fixed-order structured controllers mitigate these implementation difficulties, they inherently suffer from a performance gap compared to their full-order counterparts. This paper presents a novel hybrid tuning strategy that combines structured H∞ control with Extremum Seeking Control (ESC) to optimize disturbance attenuation in a quarter-car active suspension system, specifically targeting the H∞ norm within a critical frequency band. Initially, a structured sub-optimal H∞ controller is designed using conventional robust control methodologies. Subsequently, ESC is uniquely employed to iteratively fine-tune controller parameters, exploiting known disturbance characteristics without requiring explicit system modeling. The proposed ESC-based tuning significantly enhances the sub-optimal controller’s performance, narrowing the gap between structured and full-order optimal controllers. Stability of the closed-loop system is maintained throughout the tuning process by employing small perturbations, slow adaptation rates, and continuous evaluation of the H∞ norm. Simulation results clearly demonstrate the effectiveness of the proposed method, highlighting its potential to improve practical disturbance attenuation and overall performance.