Turbo product codes for optical communications and data storage

With the increasing demand for high-speed communications systems and high-density data storage devices, forward error correction (FEC) codes and signal processing techniques are receiving much attention. This thesis focuses on a sub-class of iterative turbo-like FEC codes called turbo product code (TPC). In contrast to the “original” convolutional turbo codes, TPCs are based on block codes, and hence are able to be decoded via algebraic methods. The first objective of this research is the reduction of the implementation complexity of the original TPC decoding algorithm. This is achieved via an efficient Chase decoder that can reduce the decoding complexity by an order of magnitude with no performance degradation. Furthermore, to decrease decoding latency, a novel parallel TPC decoding algorithm is introduced. For specific two-dimensional product codes, it is shown that parallel decoding reduces the decoding time further to half of that of a serial TPC decoder introducing only a slight loss in performance. In addition to efficient decoding methods, potential TPC application areas are also presented in this study. First, TPCs for optical fiber links and the combination of optical and/or electrical equalization techniques for intersymbol interference compensation are investigated. The second application considered is an optical code-division multiple-access (CDMA) system, for which the TPC decoding algorithm is adapted to enable more users and a reduction in required system bandwidth. In the final part of this study, the TPC performance on magnetic data storage systems is shown and further improvement possibilities are identified.