Research On Encoding And Decoding Algorithms Of LDPC Codes In CMMB System

China Mobile Multimedia Broadcasting System(CMMB), the broadcast television industry standard, is proposed by the State Administration of Radio in 2006 with independent intellectual property rights. Its core is our self-developed multimedia broadcasting transmission technology(STi Mi). CMMB-STi Mi standard uses RS codes and low density parity check(LDPC) codes to form the concatenated codes as a channel error correction coding techniques.LDPC codes have many advantages, such as simple description, low decoding complexity, can be implemented in parallel, close to the Shannon limit and have lower platform etc, It has become one of the key technologies of the fourth generation of mobile communications, and will be widely used in deep space communications,mobile and fixed wireless communications, satellite broadcasting and other digital video and audio areas, it is the most watched research hot spot in the field of channel coding. As the coding and decoding scheme for CMMB system, LDPC codes’ decoding performance will directly affect the efficiency of the whole system and its realizability. Based on the characteristics of LDPC codes in CMMB standard, this assay proposes a reasonable decoding scheme for it. The main work in this paper can be summarized as follows:1.Study the structural characteristics and construction methods of LDPC codes based on CMMB standard, CMMB standard defines two rate LDPC code, 0.5 and0.75 rate respectively. Their code length is 9216, can be constituted by the former cyclic shift sub-matrix, sized of 921618? and 92169? respectively. Based on the structural characteristics of LDPC codes in CMMB standard, make rows and columns transforms for the check matrix, the transformed parity check matrix would have a quasi-cyclic structure, which is conducive to simple coding and hardware.2.The traditional decoding algorithms, including the belief propagation algorithm(BP) and min sum algorithm(Min Sum), they are all based on the flooding message passing schedule, decoding in parallel. But The layered decoding algorithm(Layered Belief Propagation, LBP) is base on the serial message passing schedule,decoding serially, can greatly accelerate the convergence rate of iterative message. In this paper we compare the decoding performance of these two types of algorithm,under the conditions of different code lengths, code rate and maximum number of iterations, we can see that the LBP algorithm can not only speed up the message convergence rate of iterative, but also improve the decoding performance of decoding algorithms based on flooding messaging mechanism when the maximum allowable number of iterations is small. Meanwhile can save the storage space.3.In this paper presents two improved algorithms based on LBP algorithm,named grouped serial decoding algorithm(Grouped LBP) and semi-serial decoding algorithm respectively. Firstly introduce the improve ideology and the message iterative process of this two algorithms. Then simulate the decoding performance of the two improved algorithms under the condition of different code lengths, code rate and maximum number of iterations. The simulation results show that they can not only keep a good decoding performance as the LBP algorithm, but also accelerate the decoding speed. At last, use the two improved algorithms, completed the decoding simulation for the LDPC codes based on CMMB standard, and verify their realizability in CMMB system.