Hybrid Method for Convolutional Turbo Codes Decoding
Keywords
Abstract
The widespread use of error-correcting coding in modern digital communication systems is due to the fact that it is an effective method of bringing the system parameters to the desired compromise between the probability of data transmission, the required signal power, and channel bandwidth. One of the important achievements in the theory of error-correcting codes with a random structure is the development of turbo-like structures. This article considers convolutional turbo codes, which, due to their structure and specialized decoding methods with information exchange between component decoders, allow for efficient transmission of digital data in computer systems of various functional purposes. It should be noted that the basic Berrou–Glavieux–Thitimajshima method includes a number of nonlinear resource-intensive functions. That is why suboptimal turbo decoding methods are considered, which use the approximation of the correction function during the calculations of metrics and the a posteriori logarithmic ratio of the likelihood functions. It is proposed to use a hybrid method of turbo codes decoding with automatic repeat-request, which uses one of the component methods depending on the level of the noise component in the digital communication channel.
After a theoretical description of the development, a computer simulation of the error- correcting characteristics of the system with the presented turbo decoder was performed. The results of experiments for multimedia information show that the development makes it possible to achieve good corrective ability at low signal-to-noise ratios in the digital communication channel. The article can be useful for engineers, as well as designers of data transmission systems, since it allows analyzing and synthesizing turbo codes depending on the requirements set for the codec and the data transmission system as a whole.
