Decoding Difficulty Estimation of the Convolutional Turbo-Codes and Turbo-Product Codes
Keywords
Abstract
In the period of rapid development of modern digital technologies, taking into account the increase in communication range and information volumes, provision of reliable data transmission is an integral requirement for communication systems of various functional purposes (space communications, digital television, programmable radio systems, optical communications, digital data storage, etc.). In this context, turbo-codes stand out, which represent a powerful class of error correction codes with a unique codec structure, able to work efficiently at high speeds and almost totally use the capacity of communication channels. However, the effective use of turbo-codes is associated with significant computing costs at the decoding stage, which creates a significant load on the system, especially in the working "online"-mode and in case of limitations on the part of the computing platform. Understanding the impact of turbo-code parameters on their decoding computational complexity enables to provide the balance between the error correction reliability and the efficiency of using computing resources in various communication scenarios.
That is why, the article analyzes in detail the computational complexity of the decoding process of these codes. The iterative exchange probabilistic algorithms of Berrou–Glavieux–Thitimajshima, Viterbi-Hagenauer and extended Pyndiah-Chase list decoding were selected as the main decoding algorithms. In this work, generalized analytical expressions of the decoding complexity depending on several parameters of convolutional and block component codes in the case of turbo-code implementation in mixed mode using a digital signal processor are obtained, and corresponding graphical representation of the research results is given. The article is of practical importance for engineers and designers of digital data transmission systems, as it helps to analyze and synthesize more efficiently turbo-codes depending on the set requirements.
