多元LDPC随机译码的似然概率快速生成算法

Low-Complexity Probability Generation Algorithm for Stochastic Decoding of Non-binary LDPC Codes

多元随机译码算法能有效解决译码复杂度高的问题,但似然概率生成的计算仍有较高的计算复杂度.为解决这一问题,文中提出一种针对M阶正交振幅调制下的多元低密度奇偶校验码的随机译码算法.通过计算星座图中部分星座点和接收信号点之间的欧几里德距离来确定随机译码的似然概率,而不必计算所有星座点的似然概率,省去了复杂的幂次运算.仿真结果表明,该算法能有效降低多元LDPC码随机译码的计算复杂度,性能损失较小,是一种适合多元LDPC译码器前端实现的译码算法.

Stochastic decoding algorithms for non-binary LDPC codes have good performance, but high computational complexity in probability generation. In this paper, we propose a simple method for fast generation of the probability, especially in high-order modulation schemes. The algorithm calculates probability using the quantized Euclidean distances between constellation points and the received signal points. It can avoid the probability calculation for all constellation points and the arithmetic of powers. Simulation results show that the proposed algorithm has low complexity and incurs less performance degradation, which is advantageous in hardware implementation.