Clustering-based Control Optimization of Hard Disk Drive Servo System for Population-level Performance Enhancement

This paper presents a data-driven control design framework for population-level robustness and performance enhancement in hard disk drives (HDDs). First, a frequency-domain control design scheme is presented, where the stability, robustness, and performance properties of the closed-loop system are simultaneously optimized through a numerical optimization scheme. The validity of the controller is verified through a time-domain simulation. Then, a clustering-based control optimization framework is proposed to further improve the population-level performance of the system. Simulation results indicate that plant clustering based on the spectral frequency response data provides an effective solution for designing high-performance servo controllers for the HDDs.