TY - CONF

T1 - Soft binary block decoders based on a generalized Wagner rule

AU - Snyders, Jakov

AU - Be'ery, Yair

PY - 1988

Y1 - 1988

N2 - Summary form only given, as follows. According to a result of Wagner, soft decision decoding of a binary linear rate (n-1)/n block code is achieved by complementing a single, least reliable, bit of the binary sequence, which is the output of a hard limiter applied to the entries of the received sequence. A generalization of this rule for binary linear block codes with multiple (λ) check-bits is presented. Maximum-likelihood decoders with significantly reduced computational complexity are attainable for a few mid- and high-rate codes by efficient implementation of the generalized Wagner rule. The original Wagner rule, as well as the trivial soft decoding rule for rate 1 codes, were recently utilized by several authors (Conway and Sloane, Be'ery and Snyders, Forney) for deriving soft decoders with reduced computational complexity for codes that contain a subcode which is related to a single, respectively zero, check-bit code. The generalized Wagner rule is also applicable, in a similar fashion, to codes that contain a subcode related to a λ check-bit code. It is demonstrated that maximum-likelihood decoders with increased computational efficiency are obtainable, for some mid-rate codes, with the aid of the generalized rule.

AB - Summary form only given, as follows. According to a result of Wagner, soft decision decoding of a binary linear rate (n-1)/n block code is achieved by complementing a single, least reliable, bit of the binary sequence, which is the output of a hard limiter applied to the entries of the received sequence. A generalization of this rule for binary linear block codes with multiple (λ) check-bits is presented. Maximum-likelihood decoders with significantly reduced computational complexity are attainable for a few mid- and high-rate codes by efficient implementation of the generalized Wagner rule. The original Wagner rule, as well as the trivial soft decoding rule for rate 1 codes, were recently utilized by several authors (Conway and Sloane, Be'ery and Snyders, Forney) for deriving soft decoders with reduced computational complexity for codes that contain a subcode which is related to a single, respectively zero, check-bit code. The generalized Wagner rule is also applicable, in a similar fashion, to codes that contain a subcode related to a λ check-bit code. It is demonstrated that maximum-likelihood decoders with increased computational efficiency are obtainable, for some mid-rate codes, with the aid of the generalized rule.

UR - http://www.scopus.com/inward/record.url?scp=0024123857&partnerID=8YFLogxK

M3 - ???researchoutput.researchoutputtypes.contributiontoconference.paper???

AN - SCOPUS:0024123857

SP - 193

ER -