一种参量可调的多元LDPC译码算法及其能耗分析

A Parameter-adjustable Non-binary LDPC Decoding Algorithm and Its Energy Consumption Analysis

在加权比特可靠度(Weighted Bit-reliability,wBRB)多元低密度奇偶校验(Low Density ParityCheck,LDPC)译码基础上,提出了一种参量可调的译码算法。迭代过程中的比较参量不再使用固定的硬判决符号,而是基于大数逻辑准则选取最为可靠的外信息符号作为标准的比较参量,提高距离修正参数选取的准确性。在复杂度分析方面,提出了一种基于能耗的综合评判准则,将元素间的操作折算到相应的能量消耗指标上,可更加科学、直观地对不同算法的译码复杂度进行统一衡量。仿真结果显示,所提出算法的错误平层略低于原算法,其增加的能耗几乎可以忽略。

A parameter-adjustable decoding algorithm is presented based on the weighted bit-reliability (wBRB) for non-binary low density parity check(LDPC) codes. For the presented algorithm,the most reliable extrinsic information rather than the fixed hard-decision information is utilized as the standard comparison parameter based on the majority-logic criteria,which can improve the accuracy of the Hamming distance correction parameters. Furthermore,a general evaluation criterion is presented for the complexity analysis based on energy consumption. The operations between Galois field elements can be converted to the corresponding energy consumption,which can compare the decoding complexities among different algorithms more reasonably and intuitively. Simulation results show that the presented algorithm has a lower error floor than the original algorithm with negligible increasing energy consumption.