Research On Decoding Algorithm Of LDPC Codes In CMMB System

The Low Density Parity Check codes (LDPC) are a kind of the linear block codes.The difference between LDPC codes and regular linear block codes is the most of theelements in parity check matrix are0. The parity check matrix can construct lowcomplexity and good performance LDPC codes.Due to its simplie description, high throughput and good encoding and decodingperformance, LDPC codes have been widely applied to communication fields, such asMobile TV, optical fiber communications, mobile communications and fixed wirelesscommunications and audio dissemination. CMMB (China Mobile MultimediaBroadcasting) have been promulgated by The State Administration of Radio Film andTelevision of China to the mobile TV and multimedia standard. LDPC codes are alsoused as an encoding and decoding scheme in the standard.The optimization of decoding algorithm and application in real systems is theresearch hotspots of LDPC codes nowadays.A good decoding algorithm should strike abalance between its decoding performance and complexity.The length of LDPC codesin CMMB system is great. The existing decoding algorithms in CMMB system willspend a lot of hardware resources because of the high complexity.So improvingexisting decoding algorithm to reduce the computational complexity of decoding isnecessary.The decoding performance and computational complexity of the typical decodingalgorithm of LDPC codes are studied and analyzed in this thesis, such as BP(BeliefPropagation)algorithm,MS(Min Sum)algorithm and LBP(Layered Belief Propagation)algorithm.Then LBP decoding algorithm which has better decoding performance isselected as the basis for the improving algorithm.A function of f(x) is introduced in the first step to reduce the computationalcomplexity of LBP decoding algorithm, and this improved decoding algorithm is named FLBP.Although the computational complexity of LBP decoding algorithm isreduced, but the decoding performance is reduced too. Specific parameters offset factorα and correction factor β are introduced in the second step, the purpose of this step is toenhance the decoding performance which lossed in first step, and the improveddecoding algorithms in this step are named NLBP and OLBP. Suitable values of α andβ can make better decoding performance of NLBP decoding algorithm and OLBPdecoding algorithm than decoding performance of LBP decoding algorithm.In order to validate the performance of improving algorithm, the performance ofBP, MS, LBP, the improved NLBP and OLBP algorithms are simulated in this thesisused Matlab software, and the simulation curves are gotten. The performance of theimproving decoding algorithm and LBP algorithm are compared. Then theperformances of different coding algorithms with different code length and decodingiterations are compared. The simulation results show that the decoding performance ofimproved NLBP and OLBP algorithm is better than LBP algorithm. At the end of thethesis, the simulation conditions of CMMB system are set according to thecharacteristics and specific technical parameter, and the practical applicationsimulation of the improving algorithm is simulated. The result of the simulatationshow that the improving algorithm applied to CMMB system is verified feasibility andeffectiveness.