Discrete-time LQG/LTR design and modeling of a disk drive actuator tracking servo system

This paper presents the discrete-time LQG/LTR design of a disk drive track following servo system. The servo compensator designed through linear-quadratic Gaussian control combined with loop transfer recovery (LQG/LTR) consists of a Kalman filter for state estimation and state feedback for control. The desired tracking servo performance is first formulated through a frequency shaped return ratio of the Kalman filter and subsequently recovered at the output of the plant/compensator loop through the automatic design of a discrete-time linear quadratic (LQ) regulator. Particular attention has been given to modeling the calculation time delay and bias force estimation. The excellent robustness and performance characteristics of a continuous time LQG/LTR design are theoretically unachievable due to the extremely low sampling rate and nonminimum phase plant characteristics. However, both time and frequency domain simulations show that reasonable stability margins and performance can still be recovered. This technique nearly eliminates all the trial and error typical of a conventional pole placement design of a similar system. The direct discrete-time design can handle extremely low sampling rates associated with embedded servo systems. The technique can also be used for designing multi-rate and multi-input servo systems