A Method for Decoding a Variable-Rate Error-Correcting Code Based on Finite-State Automata

Abstract

A decoding algorithm for a special error-correcting code is considered and its efficiency is estimated. This code is obtained as a result of processing information messages by finite automata and using a two-base numeral system. A general encoding algorithm is also considered. Both encoding or decoding are performed by a two-level system in which an input message is represented as a lower (2,3) code at the internal level, and the error correcting capabilities of this code are strengthened owing to its transformation with the help of a special finite automaton at the external level. In decoding, errors are first detected and corrected at the external level, and then possible remaining errors are eliminated by an internal automaton. The relationship between the external level of the system being considered and convolutional codes is investigated and the advantages of the proposed method are shown.