| 引用本文: | 刘士平,马林华,胡星,黄天宇.低复杂度极化码SCL译码算法[J].哈尔滨工业大学学报,2018,50(5):160.DOI:10.11918/j.issn.0367-6234.201706116 |
| LIU Shiping,MA Linhua,HU Xing,HUANG Tianyu.Alow complexity SCL decoding algorithm for polar codes[J].Journal of Harbin Institute of Technology,2018,50(5):160.DOI:10.11918/j.issn.0367-6234.201706116 |
|
| 摘要: |
| 极化码的串行抵消列表(SCL)译码的逐次逐比特进行判决过程与路径度量值的计算筛选过程是整个译码系统复杂度与延迟的主要来源.在分析现有SCL及多比特判决译码的优缺点基础上，针对SCL译码造成的译码系统复杂度高和延时大的问题，将每组多比特码字(多个独立信道)视作一个整体，并在译码时通过信道合成构造为一个虚拟多比特信道，从而可以对多比特码字进行同步传输并根据信道递归公式进行同步判决译码.由此基于SCL译码的码树构造提出一种构造多比特虚拟信道的SCL译码算法，并结合设置译码码树节点阈值减少码树节点分裂次数的方法进一步增强了上述算法.在AWGN信道下的分别对虚拟2、4和8比特信道SCL译码的误码率及误帧率性能进行仿真.仿真结果表明在虚拟8比特信道情况下，预设阈值S=30时的译码性能接近传统SCL算法，且总历经节点数降低了63.7%，总加法次数是8比特同时判决译码算法的17%.此算法降低了译码算法的计算复杂度及硬件存储复杂度，更适合于硬件实现, 具有一定的实用价值.
|
| 关键词: 极化码 串行抵消列表译码 译码复杂度 阈值 |
| DOI：10.11918/j.issn.0367-6234.201706116 |
| 分类号:TN911.22 |
| 文献标识码:A |
| 基金项目:国家自然科学基金资助项目(61472442); 综合业务网国家重点实验室(西安电子科技大学)开放研究课题资助项目(INS-15-13); 航空科学基金资助项目(20155896025) |
|
| Alow complexity SCL decoding algorithm for polar codes |
|
LIU Shiping,MA Linhua,HU Xing,HUANG Tianyu
|
|
(Institute of Aeronautics and Astronautics Engineering, Air Force Engineering University, Xi’an 710038, China)
|
| Abstract: |
| The successive cancellation list (SCL) decoding algorithm of polar code is of bitwise, which is the main source of the decoding complexity and latency. In view of the analysis of the advantages and disadvantages of the existing SCL and multi-bit decision decoding, each set of multi-bit code words (multiple independent channels) is treated as a whole to construct a multi-bit virtual channel, which can transmit multi-bit code words and decode synchronously. Therefore a construction of multi-bit virtual channel based on the SCL algorithm under the AWGN channel is proposed to against the problem of high complexity and large delay for decoding system caused by SCL decoding. It is also enhanced by combining with the establishment threshold of the decoder tree nodes to reduce the number of tree nodes' split. Under the AWGN channel, the bit error rate and the frame error rate performance of the virtual 2, 4 and 8 bit channels are simulated. The simulation results show that the decoding performance is close to that of the traditional SCL algorithm when the default threshold S=30 in the case of a virtual 8-bit channel. The total number of nodes is reduced by 63.7% and the total number of additions is 17% of the 8-bit decision decoding algorithm. The proposed algorithm can effectively reduce the calculation complexity of decoding and the hardware storage. It is more suitable for hardware implementation and has a certain practical value.
|
| Key words: polar code successive cancellation list algorithm decoding complexity threshold |
