Weight-Based Binary Channel Coding—A Machine Learning Approach

Abstract

AbstractChannel encoding provides the ability of having a reliable error-free reception in connection with digital data transmission. The channel noise affects some of the information bits transmitted (i.e., the received and the transmitted will not be the same), and such affected bits are referred to as errors in the received. Channel coding attends such influences of additive channel noise on the information transmitted, through the redundancy provided for the information prior to transmission. This redundancy provides overheads for transmission, and these are responsible for error-detecting/-correcting features of the communication system. The error-detecting/-correcting ability of a particular channel coding scheme depends on its Hamming distance. In the present discussion, the encoding and decoding (machine learning approach) of single error-correcting binary channel coding scheme referred to as weight-based code is proposed. An \(^{\prime}n^{\prime}\) bit code word will be transmitted in the place of \(^{\prime}k^{\prime}\) bit Information word, where \(n = k + r\), (modulo-2 sum) is obtained by appending \(^{\prime}r^{\prime}\) number of parity bits to \(^{\prime}k^{\prime}\) bit Information word. Code efficiency is a measure of the useful number of bits in the total number of bits transmitted. Weight-based code is such \(\left( {n, k} \right)\) binary coding scheme with a code efficiency of \(\frac{k}{n} \times 100\% = 50\%\) and is a systematic single error-correcting code.KeywordsWeight Based CodeParity WordCode efficiencySystematic codeHamming Distance