D.A. Kapustin, A.A. Gladkikh, V.E. Dementew, V.V. Voronin

Modern ICT is organized as the interconnected system of information exchange based on integrating communication systems including ground and satellite, cellular and fiber-optic communication lines and their development by means of an elemental base of new generation. A considerable proportion of such communication systems include channels, where it is impossible to ensure high energy potential. One of the solutions of achieving required reliability characteristics in such systems is applying noiseless coding. Most of nowadays investigations in this field are connected with systems of turbo coding using methods of soft-decision decoding by means of iterative code combination transformations. In a soft-decision decoder each bit of information is followed by a symbol-s accuracy index, which is sometimes defined as symbol-s reliability evaluation. There are different approaches to forming a symbol-s accuracy index. It can be performed on the basis of log-likelihood ratio, quantization of modulated signal parameter into several levels, etc. The main disadvantage of each method is the dependency of a symbol-s accuracy index from parameters of a communication channel. One proposes the method of forming a symbol-s accuracy index on the basis of an erasure channel with a long erasure interval. In this case signals registered close to the mathematical expectation have the highest symbol-s accuracy index. Erasures differ in reliability according to a linear function determined by a receiver constructor. Using a symbol-s accuracy index enables to decode block codes outside the limits of their facilities. One shows energy advantage of applying the method based on decoding equivalent codes using a symbol-s accuracy index.