Honda, Masanori2016-12-192016-12-192016-10https://hdl.handle.net/11299/183370University of Minnesota Ph.D. dissertation. October 2016. Major: Aerospace Engineering and Mechanics. Advisor: Peter Seiler. 1 computer file (PDF); ix, 117 pages.A hard disk drive (HDD) is a device that stores digital data by writing and reading magnetic signals onto a disk using a magnetic transducer. The data is organized into circular tracks that are less than 100 nm wide. Modern HDDs use a dual-state actuator to position the transducer onto these tracks. Since the tracks are extremely narrow, a high performance controller is required to reject disturbances from various sources and maintain the position of the transducer on a single track. This dissertation focuses on designing a robust controller for the HDD system. The controller must be robust to its external environment, such as changes in temperature, and provide good performance to thousands of drives. An adequate uncertainty model designed using first principles is not available for robust controller design. Thus a set of frequency response data (FRD) measured from a number of HDDs and at different temperature points is used to design the uncertainty model of the system. A basic method of averaging the set of FRD to create an uncertainty model is used to design the baseline controller through a standard D-K synthesis method. A numerical algorithm is then developed to create an optimal uncertainty model for the system using the experimental FRD. Using this algorithm, a temperature dependent model is designed for the purpose of designing a temperature dependent robust controller. Finally a temperature dependent controller is designed to increase the performance of the HDDs compared to the baseline controller, and the theoretical validation for the method is given.enHard Disk DriveParameter VaryingRobust ControlsThesis or Dissertation