Model Predictive Control for Track Following and Disturbance Rejection in a Tape Drive System

We provide Model Predictive Control (MPC) formulations for the advanced control of high-precision track-following servo systems to improve rejection of noisy, measurable disturbances. This work develops a disturbance prediction method using the wavelet denoising technique and system identification, which is used to anticipate the disturbances in the system using preview measurements. The MPC formulations allow for improved track following and disturbance rejection when compared to classical control techniques. We consider ℓ₁ and ℓ₂ formulations of the MPC problem to assess and compare tracking performance and computational complexity as a function of the prediction horizon. These advanced control techniques are simulated on a reel-to-reel tape drive system model to demonstrate the ability of these algorithms to perform precise track following and disturbance rejection while assessing implementation techniques that can yield fast computation times that this application demands.