Research On Low-complexity Decoding Algorithm For Turbo Product Code

Channel coding technology can improve the accuracy of the information transmission and Turbo Product Code(TPC) is a brilliant code with great theoretical and practical value. TPC hard decision decoding needs sample computation; as a result, it also brings poor performance. To the contrary, algorithm of soft decision decoding can gain excellent performance at the cost of high complexity. Nowadays new requirements about channel encoding and decoding have been put forward because of higher and higher information transmission rate. It is an urgent need to establish a complete and efficient theory on TPC decoding algorithm.TPC traditional hard and soft decision decoding algorithms are discussed first in this paper. As for hard cascade decoding algorithm, it’s explained that the existence of some particular error patterns that can’t be corrected leads to the poor performance. To soft Chase-Pyndiah algorithm, we understood its principle and main parameters; meanwhile, it’s clear that the number of hard decision decodings and arithmetic operations is main source and evaluation index of complexity.In the research on the optimization of hard decision decoding, the existing non-sequential algorithm is introduced first and then the concatenated decoding algorithm is proposed based on concluding some particular error patterns of original cascade decoding algorithm. As an optimization and amelioration strategy, the proposed algorithm needs only a little more computation, but performance can be effectively improved. So the concatenated decoding is a preferable algorithm with good performance and low complexity. It can be applied to some systems where real-time communication is needed or soft information is difficult to obtain as a powerful supplement of soft decision decoding algorithm.Traditional soft decision Chase-Pyndiah algorithm is the mainstream of TPC decoding. Part of the study to simplify the algorithm is concerned about the details such as Euclidean distance calculation, the structure of decoding and the competition codeword finding and so on. We optimize the traditional algorithm from the macroscopic aspects in this paper and then the formula method and the new adaptive decoding algorithm are proposed. Formula method can select way to extrinsic information calculation according to the reliability of row or column codeword and achieve a significant reduction in complexity with no loss of performance through simplifying the process. The new adaptive decoding algorithm put the idea of formula method to the original adaptive algorithm. With iteration process going, the number of unreliable bits can be reduced adaptively; at the same time, extrinsic information calculation of reliable codeword can also be simplified. The performance of the new adaptive algorithm falls slightly; nevertheless, a sharp drop in complexity can be gained compared with the traditional algorithm. The proposed adaptive algorithm can be applied to communication system where performance requirement isn’t very strict. In a word, both formula method and the new adaptive algorithm can be used as effective methods to TPC soft decision decoding.