Estimation of tracking errors using array heads

The general theory of array heads for data recovery and servoing purposes is described for both decision-oriented and nondecision-oriented approaches. Simulation results for both these approaches using a two-track, three-head model are shown, demonstrating the convergence of the estimate of the misregistration, in , as a function of the SNR. The decision- and nondecision-directed approaches appear to provide similar estimates of in for the 1D channel. However, the nondecision-directed approach can be computed independently of the data detection process, if one has knowledge of the noise correlation, making this method simpler to implement.<>