针对卷积LDPC码译码时延长的问题,提出了一种高效的译码算法.在每步迭代过程中运用反馈消息,能更加有效地更新变量节点消息,并采用比重因子法减小了由于因子图中环的存在而产生的误差传播,从而大大减少了译码迭代次数,提高了译码的收敛速度.仿真结果表明,该译码算法减小了5/8的译码时延,并降低了译码复杂度,同时获得了比现有的置信传播算法更好的纠错性能,而且在相同的迭代次数下,本算法在BER为10^-6时获得了0.16 dB的增益.
A novel belief propagation (BP) decoding algorithm for convolutional low-density paritycheck codes is proposed. The proposed algorithm raises the efficiency of updating the variable information by applying feedback information at each decoding iteration and employs the weighting factor to reduce the error propagation caused by the cycles in the Tanner graph, thus yielding a faster convergence of the decoding. Simulation results show that an error performance better than that of the existing belief propagation algorithm can be achieved, while the 5/8 decoding delay and the computation complexity are effectively reduced. Compared with the existing BP, the proposed algorithm achieves a gain of 0. 16 dB at the BER of 10^-6 with the same number of iterations.