Erozan Kurtas

Multi-Track Matched Spectral Null Codes for the Partial Response Maximum Likelihood Magnetic Recording Channels

Abstract

In this thesis, we investigate the problem of designing matched-spectral null (MSN) codes for multi-track magnetic recording systems. The ultimate goal of this research is to design high rate, high gain multi-track MSN codes that will result in an overall increase in data storage density. We give an analysis of multi-track systems which allows us to identify the major sources of system degradation and develop two-track and three-track high rate MSN codes. We show that these codes are able to deal with intersymbol interference and inter-track interference very effectively. Moreover, we provide results for multi-track systems with servo positioning errors and investigate the performance of the codes developed in this work. The last problem we attack in this dissertation is to design reduced complexity detection algorithms which show almost optimal performance while resulting in a low complexity for implementation purposes.