Decision-feedback maximum-likelihood decoder for finite-state Markov channels

Andrea Goldsmith, Pravin Varaiya

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Scopus citations

Abstract

Channels that are time-varying due to fading, interference, or variable noise can be modeled as finite-state Markov chains. These channels have memory due to correlated variations, which should be incorporated into the design of an optimal code. However, optimal codes of this type are difficult to design, and the complexity of their decoding algorithms grows exponentially with memory length. Thus, memoryless codes are typically used and the encoded bit stream is interleaved prior to transmission to remove the effect of memory. This results in a performance loss relative to the inherent channel capacity [1]. In this paper, we develop a decision-feedback decoding algorithm that uses the channel's Markovian structure to determine the maximum-likelihood input sequence. We show that this decoding scheme can achieve data rates approaching the Shannon limit with a moderate complexity increase over the conventional approach. We also present numerical results for the capacity and cutoff rate of a two-state fading channel with 4PSK modulation using both the decision-feedback decoder and the conventional memoryless encoding method.

Original languageEnglish (US)
Title of host publicationConference Record - International Conference on Communications
PublisherPubl by IEEE
Pages1546-1550
Number of pages5
ISBN (Print)0780318269
StatePublished - 1994
Externally publishedYes
EventProceedings of the 1994 IEEE International Conference on Communications - New Orleans, LA, USA
Duration: May 1 1994May 5 1994

Publication series

NameConference Record - International Conference on Communications
Volume3
ISSN (Print)0536-1486

Conference

ConferenceProceedings of the 1994 IEEE International Conference on Communications
CityNew Orleans, LA, USA
Period5/1/945/5/94

All Science Journal Classification (ASJC) codes

  • Computer Networks and Communications
  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Decision-feedback maximum-likelihood decoder for finite-state Markov channels'. Together they form a unique fingerprint.

Cite this