极化码BP译码算法中量化问题的研究

Exploration on Quantification Problems for Polar Code with Belief Propagation Decoder Algorithm

第五代移动通信(5G)面向高可靠和低时延的海量数据交互场景,如何实现其高效编译码方案是一个亟需解决的问题。现有编译码方案难以有效平衡译码过程中的计算复杂度和硬件实现难度,故无法应用于5G网络架构。针对该问题,提出了一种基于极化码的置信度传播(BP)译码量化方案。该方案结合最小和译码的近似算法和量化思想,显著降低了BP译码算法的计算复杂度和硬件实现难度。仿真成果表明,与最小和译码算法相比,量化后的BP译码算法极大地提高了译码性能。相比于BP译码算法,在几乎不降低译码性能的基础上,量化后的BP译码算法明显降低了计算复杂度,更便于硬件实现。

5G mobile communication orients itself to the high-reliability and low-latency interaction scenarios of mass data, how to realize high- efficiency coding/decoding is a problem demanding prompt solution, and however, the present coding/decoding scheme is difficult to effectively balance the computational complexity and the hardware realization in decoding process. For this reason, the belief key issues are how to balance the computational complexity and hardware realization of the decoding process in 5G system. Therefore, the belief propagation decoding quantification scheme based on polar code is proposed. This scheme, with approximate algorithm of min- sum decoding and quantitative idea, could remarkably reduce the computational complexity and difficulty of hardware implementation. The simulation results indicate that the quantized belief propagation decoding algorithm could greatly improve the decoding performance as compared with the min-sum decoding algorithm. Compared with the belief propagation decoding algorithm, this proposed algorithm could obviously reduce the computational complexity and facilitate the realization of hardware with little degradation of decoding performance.