CMMB标准中LDPC码的译码及定点仿真研究

Study of the LDPC Decoding and the Fixed-point Simulation in CMMB

针对CMMB标准中的LDPC码进行了译码算法的研究,提出最小的改进算法,同时给出归一化最小和算法的定点仿真。定点仿真结果表明,输入8位量化,中间变量分别9位和10位量化,可获得与浮点相差很小的误码率性能。

In this paper, we research the decoding algorithms of the LDPC code in CMMB and propose an im- proved algorithm based on the Min-Sum algorithm. The simulation results indicate that the performance of the improved Min-Sum is much better than the MS algorithm and has a little degradation compared with the NMS algorithm in the case of relatively high SNR. And the improved algorithm is easier to be implemented. Moreover, the fixed-point simulation of the normalized rain-sum algorithm is also studied. The result shows that the BER performance only has a slight decline with the input 8 bit quantized and inner variable respectively 9 bit and 10 bit quantized. 6 bit quantization of the input and 8 bit quantization of the inner variable is also acceptable when the memory is sensitive.