The reliability measure for a decoded symbol is the probability Pc that the symbol is correct or the probability of error Pe=1-Pc. Such quantities can be obtained by thesymbol-by-symbol MAP (maximum a posteriori probability) algorithm.Unfortunately this algorithm is computationally inefficient. A softoutput Viterbi algorithm (SOVA) can provide an estimate of Pewhich is accurate only for large SNR. This paper proposes an efficientmodified MAP algorithm for obtaining Pc for the outputs ofconvolutional inner decoders. The outer decoder uses Pc toperform soft decision decoding by choosing a codeword which maximizesthe maximum likelihood (ML) metric. Decoding based on this ML metric isreferred to as generalised soft decision decoding since it includes theEuclidean metric on AWGN channels and binary memoryless channels asspecial cases
展开▼
机译:解码符号的可靠性度量是概率P
c sub>表示符号正确或出现错误的概率为P
e sub> = 1-P c sub>。这样的数量可以通过
逐符号MAP(最大后验概率)算法。
不幸的是,该算法在计算上效率低下。柔软的
输出维特比算法(SOVA)可以提供P e sub>的估计
仅对于大SNR才是准确的。本文提出了一种有效的方法
改进的MAP算法,以获取P c sub>的输出
卷积内部解码器。外部解码器使用P c sub>
通过选择一个最大化的码字来执行软判决解码
最大似然(ML)指标。基于此ML指标的解码是
之所以称为通用软判决解码,是因为它包括
AWGN通道和二进制无内存通道上的欧氏度量为
特别案例
展开▼