A maximum likelihood decoding algorithm for turbo codes

Author

Sadowsky, John S.

Author_Institution

Dept. of Electr. Eng., Arizona State Univ., Tempe, AZ, USA

Volume

2

fYear

1997

fDate

3-8 Nov 1997

Firstpage

929

Abstract

A maximum likelihood (ML) decoding algorithm is presented for decoding turbo codes. The basic algorithm, called the pair list algorithm (PLA) is a stack algorithm that examines pairs of constituent code trellis paths in order of decreasing metric sum. The PLA controls two per trellis list Viterbi algorithms (LVA) for finding the n highest metric paths in a trellis diagram. For practical implementation the ML search must be terminated. This search truncation may limit the utility of this algorithm to ℰ_b/N₀ for which the cutoff rate R₀ exceed the code rate. However, the terminated ML search can be used to generate FER statistics, or for a FEC/ARQ turbo code scheme without using an additional error detecting outer code

Keywords

Viterbi decoding; automatic repeat request; concatenated codes; convolutional codes; forward error correction; interleaved codes; maximum likelihood decoding; FEC/ARQ turbo code scheme; FER statistics; constituent code trellis paths; decreasing metric sum; highest metric paths; list Viterbi algorithms; maximum likelihood decoding algorithm; pair list algorithm; search truncation; stack algorithm; turbo codes; Automatic repeat request; Convolutional codes; Error analysis; Forward error correction; Iterative decoding; Maximum likelihood decoding; Maximum likelihood detection; Programmable logic arrays; Turbo codes; Viterbi algorithm;

fLanguage

English

Publisher

ieee

Conference_Titel

Global Telecommunications Conference, 1997. GLOBECOM '97., IEEE

Conference_Location

Phoenix, AZ

Print_ISBN

0-7803-4198-8

Type

conf

DOI

10.1109/GLOCOM.1997.638462

Filename

638462