论文部分内容阅读
为了降低低密度奇偶检验码的误码平底,提出一种基于陷阱集状态检测的两级置信度传播译码算法。该算法分2步进行,首先,按照标准的置信度传播算法进行迭代,直到检测到陷阱集状态的出现;然后,尝试对可能出现错误的比特的初始值进行翻转扩张继续进行迭代,直到译码成功或达到最大尝试次数。仿真结果表明:该算法大大降低了误码平底区域迭代译码所需的最大迭代次数,同时在进行很少次的尝试后,低密度奇偶检验码的误码平底可以降低近2个数量级。
In order to reduce the error flatness of low-density parity-check codes, a two-level confidence-propagation decoding algorithm based on trap-set state detection is proposed. The algorithm is divided into two steps. First, the algorithm is iterated according to a standard belief propagation algorithm until the occurrence of a trap set state is detected. Then, the initial value of a bit that may have errors is flipped to continue iteration until decoding Succeeded or reached the maximum number of attempts. The simulation results show that the proposed algorithm can greatly reduce the maximum number of iterations required for iterative decoding in the flat bottom of the error-prone area, and the number of low-density parity-check codes can be reduced by nearly two orders of magnitude after a few attempts.