关键词: 信道极化码, 置信度传播算法, 等误差, 线性近似, 计算复杂度

Abstract:

In the decoding algorithm for Polar codes, the belief propagation (BP) decoding algorithm in the log-likelihood ratio domain incurs high computation complexity due to the computation of the hyperbolic functions Motivated by this observation we propose an improved BP decoding algorithm. In the node update rules, our method replaces the hyperbolic functions with the linear approximation functions based on the principle of equal error. Compared with the original BP decoding algorithm, the modified BP decoding algorithm is only implemented by addition and multiplication operations, which greatly reduces computation complexity, and simplifies hardware implementation. Simulation results show that the performance of the modified BP decoding algorithm is almost the same as that of the original BP decoding algorithm in the low Signal to Noise Ratio (SNR) region, and in the high SNR region the performance of our method is slightly worse. Compared with the original BP decoding algorithm, the bit error rate (BER) performance of the modified BP decoding algorithm has about 0.1dB degradation when the length of Polar codes is 256 and the BER is 10^-6.

Key words: Polar codes, belief propagation(BP) algorithm, equal error, linear approximation, computation complexity