Block decoding with soft output information

Abstract

A method for decoding a linear block code of information bits transmitted across a signal channel, comprising the steps of: a) receiving the transmitted signals from the channel and converting the received signals into a number of received vectors r with each received vector having a length of n; and for each vector r, b) performing hard decisions on r to produce a hard decision data vector b and a corresponding reliability information vector a and performing hard decision decoding on b to produce an error pattern e, and further converting each vector b into a center code word; the vector a and the center code word each having a length n; c) identifying a set of j=0 to p nearby code words including the center code word; d) determining, for each nearby code word a difference metric DM.sub.j, wherein each difference metric DM.sub.j is defined by the following relationship: DM.sub.j =.SIGMA..delta..sub.j (I).alpha.(I) for I=0 to n where .delta..sub.j .epsilon.(0, 1) is the Hamming difference between b and the j.sup.th nearby code word; e) identifying a minimum difference metric D.sub.min, and determining an output code word as being that code word which corresponds to D.sub.min ; and f) assigning a confidence measure for each bit m in the output code word based on a difference of difference metric values for the dominant code word with a "1" in the m.sup.th position and the dominant code word with a "0" in the m.sup.th position.

Dependencies are inferred from the claim text.