Адаптивное декодирование алгебраических сверточных кодов перемежения

Abstract

RU: Обоснована целесообразность применения алгебраических сверточных кодов перемежения для увеличения достоверности передачи информации в каналах с памятью. Предложен метод итеративного декодирования данных кодов на основе комбинированного подхода с использованием процедур адаптивного распространения доверия и природных вычислений. Представлены основные принципы реализации основных этапов предложенного метода декодирования.

EN: For correcting burst errors in channels with memory are widely used various code constructions together with interleaving. For increasing the reliability of information transmission in such channels is advisable to apply algebraic interleaved convolutional codes. The principles of polynomial and matrix representation of algebraic interleaved convolutional codes with predetermined coding rate are given. The parameters of these codes which completely determined by the modified generator polynomial are presented. Algebraic decoding these codes has been a large computational complexity and limited correction capability. Algebraic interleaved convolutional codes can be represented as a long binary block codes which can be determined with using generator or parity check matrix. The combination approach to iterative decoding these codes whose purpose is to find the most probable codeword by assessing a log likelihood ratio of each code symbol is proposed. The share using information about the reliability of symbols, adaptive belief propagation and natural computing procedures are offered. According to this approach each iteration of decoding starts with finding the expected codeword using natural computing procedures. If the received vector is not codeword, then adaptive belief propagation procedures are used for updating the information about the reliability of symbols which then used in the next iteration of decoding. The basic principles of implementation of the main steps of the proposed method decoding algebraic interleaving convolution codes are presented.