Research And Implementation Of LDPC Encoding And Decoding Algorithms In DVB-C2Standard

The European DVB-C2(The Second-Generation Digital Video Broadcasting for Cable Systems) standard has been launched in2009. Compared with DVB-C adopting signal-carrier modulation with RS coding, DVB-C2adopts orthogonal frequency division multiplexing (OFDM) with higher order QAM modulation and new channel coding (the concatenation of BCH and LDPC). These technologies greatly improve the spectrum utilization, enhance the capability to fight against harsh environments and increase the system capacity at least30%.Low-Density-Parity-Check code (LDPC) is a class of linear block code which can be characterized by a sparse parity-check matrix or Tanner graph. Their error-rate performance approaches Shannon limits and their decoding complexity is relatively low. In this thesis, we study the encoding and decoding algorithms of LDPC based on DVB-C2standard. The main contents include the following aspects:(1) The hard-decision and soft-decision decoding algorithms of LDPC are analyzed in detail. Belief Propagation (BP) algorithm in log-domain and its simplified algorithms are derived from BP algorithm in probability domain.(2) After analyzing the difference between the flooding message-passing schedule and serial message-passing schedule, we propose the Log-BP algorithms and its simplified algorithms based on serial message-passing schedule.(3) After introducing how to calculate Log-likelihood ratio in AWGN channel, we show the simulation of LDPC decoding performance in different QAM modulation. The simulation results show that serial decoding algorithms have a higher convergence rate than flooding decoding algorithms, and that the modified BP-Based algorithms achieve good tradeoff between decoding complexity and decoding performance.(4) For storing parity check matrix, we analyze the encoding algorithm of LDPC in detail. The results show that the line-based parity check matrix has quasi-cyclic structure. Therefore, we just need to store the first line of each sub-matrix, which greatly reduces the storage of the address in the decoding process. (5) We finish fixed-point implementations of the encoding and decoding algorithms in Code Composer Studio. By adopting improved encoding and decoding algorithms, combining C code and linear assembly, taking advantage of optimization techniques and feedback information which CCS compiler provides, we optimize the entire code. The execution speed of optimized code is greatly improved and the throughput fully meets the requirements of the system.