Error Control of Generalized Concatenated Systems - Communications, Computers and Signal Processing, 1997. '10 Years PACRIM 1987-1997 - Networking th

Jan Bajcsy and Hisashi Kobayashi Department of Electrical Engineering Princeton University Princeton, NJ 08544 e-mail: [email protected], [email protected] Fax: (609) 258-3745, Tel: (609) 258-1984 We study error correcting codes, which are constructed by using “permutation” and “iterative decoding” in conjunction with “concatenation”. These codes include “Concatenated codes,” “product codes” and recently invented “Turbo codes’’ as special cases. We then extend our approach to a continuous channel, by adopting AZD (ambiguity zone detection) at the receiver, which introduces generalized erasures. Iterative decoding can significantly improve the performance of a system that contains a channel with memory (e.g., spectral shaping, ISI, multipath delay) and/or modulation with memory (e.g., trellis coded modulation, continuous phase modulation), is implementable with low decoding complexity, and requires a small decoding delay.