Research And Design Of Multi-standard LDPC Decodingalgorithm

As an error control code and because of its proximity to the Shannon limitperformance, LDPC(Low Density Parity Check) code has been widely used in variouswireless communication standards including China’s terrestrial digital TVtransmission standard DTMB, Europe’s second-generation satellite digital videobroadcasting standard DVB-T2/C2/S2, IEEE802.11n, IEEE802.16e and etc. Theincreasing application of LDPC codes, making multi-standard integration LDPCdecoder design has become a trend. This article is focused on researching inmulti-standard LDPC decoding algorithm in order to lay a solid foundation for notonly supporting different code lengths or code rate LDPC decoder, but also supportingthe structure of LDPC codes under different standards.This paper focus on the decoding algorithm which can be used in differentdigital-TV standards, including DVB-T2/C2/S2and DTMB etc. My work mainlyincludes the following several parts:Firstly，this paper solved the data conflict problem when TDMP decodingalgorithm is used in DVB-T2/C2/S2LDPC codes. Based on the structure ofLDPC codes is analyzed under the different standards, this paper first adopted amethod of matrix transformation greatly reduce conflict matrix. And then, to avoid thedata collision problem completely, this paper improved the TDMP algorithm, add ingprocessing parts of the remaining DDSM which will cause conflict problem.Thereby, TDMP algorithm is also applied in the DVB-T2/C2/S2and DTMBstandard, compared with the traditional TPMP algorithm, not only greatly improveperformance, but also lower decoding complexity.Secondly, another focus of this paper is the error floor problem. First Itheoretically analyses the causes of error floor problem, and then according to thechanging of the node information, puts forward a simple algorithm, effectively reducethe error flat layer, and verified by simulation.Thirdly, in order to make the multi-standard integration algorithm can be usedfor hardware implementation, this paper also design the fixed-point quantization. Todo this, firstly change the float-point algorithm into fix-point algorithm, and thendetermine the width and precision of each node to build bridges to the final hardwaredesign. Finally, compare the performance between the float-point algorithm andfix-point algorithm, to make sure that the design meets the requirements. The multi-standard LDPC decoding algorithm proposed in this paper firstlycombine the matrix transformation with algorithm improvement to apply TDMPdecoding algorithm to DVB-T2/C2/S2and DTMB. Compared with the traditionalTPMP algorithm which is widely used in integration, the new algorithm proposed inthis paper has the advantages of superior performance and lower complexity. Inaddition, to solve the error floor problem, this paper put forwards an easy implementsolution. All of this makes the whole algorithm has a high practical value, and thefixed-point quantization also ready for hardware implementation...