Turbo Codes and Turbo Principle
Keith Q.T. Zhang
Abstract
Keith Q.T. Zhang
Abstract
This chapter discusses the generation of turbo codes which can be implemented by either serial or parallel concatenation. It presents parallel concatenation for illustration. Maximum a posteriori (MAP) decoding of turbo codes relies on a recursive algorithm that exploits the inherent trellis structures of convolutional codes. Such an algorithm was derived by Bahl, Cocke, Jelinek, and Raviv, and thus has the name BCJR algorithm. To decode a turbo code formed by parallel concatenation of two recursive systematic convolutional (RSC) codes, two constituent decoders perform the BCJR algorithm individually, but exchanging their estimation information to expedite the iteration convergence. The extrinsic information transfer (EXIT) chart intuitively describes the convergence behavior of extrinsic information exchange between the two constituent decoders, thereby providing a guideline for turbo design. The chapter investigates turbo equalization, statistics of log likelihood ratio (LLR) and application of the turbo principle such as turbo CDMA and interleaver-division multiple-access (IDMA).
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
This chapter discusses the generation of turbo codes which can be implemented by either serial or parallel concatenation. It presents parallel concatenation for illustration. Maximum a posteriori (MAP) decoding of turbo codes relies on a recursive algorithm that exploits the inherent trellis structures of convolutional codes. Such an algorithm was derived by Bahl, Cocke, Jelinek, and Raviv, and thus has the name BCJR algorithm. To decode a turbo code formed by parallel concatenation of two recursive systematic convolutional (RSC) codes, two constituent decoders perform the BCJR algorithm individually, but exchanging their estimation information to expedite the iteration convergence. The extrinsic information transfer (EXIT) chart intuitively describes the convergence behavior of extrinsic information exchange between the two constituent decoders, thereby providing a guideline for turbo design. The chapter investigates turbo equalization, statistics of log likelihood ratio (LLR) and application of the turbo principle such as turbo CDMA and interleaver-division multiple-access (IDMA).
Key concepts: Turbo code, BCJR algorithm, Convolutional code, Serial concatenated convolutional codes, Computer science, Turbo equalizer, Concatenated error correction code, Algorithm