The Study Of The LDPC Code In DVB-S.2 Standard

DVB-S.2 is an new generation of satellite TV standards published by DVB organization in June 2004.It adopts large innovations mainly in three aspects:using LDPC and BCH concatenated codeing as the error correction encoding,8PSK,16APSK,32APSK as the modulation mode and VCM combined with ACM as the work pattern.Compared with DVB-S,the DVB-S.2 has improved markedly in the bandwidth efficiency (boosts bandwidth by 35% under the same power dissipation) and the use of the LDPC and BCH concatenated codeing has played a critical role.Low-Density Parity Check Codes(LDPC codes),named for the majority 0 and minority 1 in its check nodes, is a kind of linear block code with sparse parity check matrix,and a class of linear block error correcting codes defined with sparse parity check matrix. This check matirx has following characteristics:Possesses better performance than Turbo Codes in many cases, more flexibility and lower error floor character, briefly described therefore makes strict theoretical analysis verifiable, lower decoding complexity compared to the Turbo Codes and completely parallel algorithm to guarantee extremely high speed encoding and reduced hardware complexity,hence, it is suitable for hardware implementation. Meanwhile, attributes to the great throughput,there is scope for high-speed decoding.LDPC codes, with the performance of approaching the Shannon limit, has good application prospects. In the standard of DVB-S.2, LDPC codes, working as internal codes, cascades with BCH codes which work as external codes to the opposite, making the correcting ability has been improved further. In this thesis the characteristics of LDPC under the standard of DVB-S.2 are discussed in the following aspects:1.Established a current encoding and decoding simulation system in LDPC. Construct module with sparse matrix and create check matrix, signal source module produces uniformed distributed source sequences and output to encoding module to create LPCD codewords. Mutiple nosie channel models with adjustable parameters were added to channel module.Decoding module reduce codewords that has been interfered by channel noises to the original signal source information based on the decoding arithmetic applied by the interior.Finally check module will calculate the bit error rate. Simulation system possesses sparse matrix module created by MacKey stucturing method, RA,RU coding algorithm,BP/LLR-BP,minimum summation decoding algorithm and special sparse matrix,coding algorithm,decoding algorithm interface. The simulation system is also capable of searching for experimental points automatically, simulating the minimum experimental points needed when drawing SNR-BER curves rapidly. Simultaneously,by programming the codes with C language and using modular design method, we can achieve high effecienct and accuacy simulation, have a high degree of portability and realize by hardware only by modifiying function a little.2.Researched on the stucture and coding algorithm of LDPC codes in the standard of DVB-S.2.Based on the cycle accumulation, the disadvatages caused by the encoding part with greater complexity and the large storage can be solved. A LDPC encoder with linear coding complexity could be devised using the principle mentioned above. Credited to the special coding algorithm,only a comparatively less coding table file is needed to store instead of generating matrix,which make it convient for the hardware of fast encoder.3.The Gallgaer probability decoding algorithm,BP/LLR-BP algorithm,minimum summation algotithm and its improvement is considered. A new type of LDPC decoding algorithm with low complexity is propounded on them.After initializing info node, we should iterative update check node, when itertation check node returned to information code after obtaining message processing from adjacency.In this process,the key point, hyperbolic tangent function whitch is a kind of nonlinear operation was replaced by sectionalized piecewised table-look-up function, which reduced the algorithm complexity on a large scale. At last, a LDPC decoder was achieved based on this algorithm. The result of the simulation showed that this algorithm could obtain high decoding quality (only< 0.05dB less then BP algotithm performance when iterated 50 times). This decoder demands less in CPU word length in operation platform and arithmetic speed and has higer practical value.