一种改进的LLR BP译码算法研究

摘要

运用LLR BP经典算法对低密度奇偶校验(LDPC)码译码时,由于译码时迭代次数过多和每次循环时校验节点的计算复杂度过高,导致译码复杂度非常高.提出了一种改进型LLR BP译码算法,采用泰勒级数将LLR BP算法中复杂度高的雅克比修正项进行分段线性近似.仿真表明:该算法在译码性能损失不大的情况下可大幅降低LDPC码的译码复杂度.%When classical log likelihood rate belief propagation(LLR BP)is applied to decode low density parity check(LDPC) code,it becomes very complex because of excessive number of iterations and high computation check complexity of parity check nodes at each iteration.To solve this problem,a modified LLR BP decoding algorithm is proposed,which approximating the highly complex Jacobi correction term in LLR BP algorithm segmentally and linearly by Taylor Series.Simulation shows that it becomes less complex than LDPC decoding in case of not too much loss of decoding performance.