基于均衡约束的ADMM-LDPC译码算法  

ADMM-LDPC decoding algorithm based on equilibrium constraint

作　　者：夏巧桥 张迪科 王泉德[2] Xia Qiaoqiao;Zhang Dike;Wang Quande(College of Physical Science and Technology,Central China Normal University,Wuhan 430079,China;School of Electronic Information,Wuhan University,Wuhan 430072,China)

机构地区：[1]华中师范大学物理科学与技术学院,湖北武汉430079 [2]武汉大学电子信息学院,湖北武汉430072

出　　处：《华中科技大学学报（自然科学版）》2018年第8期77-82,88,共7页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61501334);中央高校基本科研业务费专项资金资助项目(CCNU16A05028)

摘　　要：基于交替方向乘子法（ADMM）的改进型惩罚函数LDPC译码算法能够提升译码性能,但却存在优化参数数目过多的问题.针对该问题,提出一种基于均衡约束的ADMM-LDPC译码算法.首先将码字的0-1整数约束等价转化为连续的均衡约束,使得原有的整数优化问题转化为含均衡约束的非凸问题;然后采用惩罚函数的方法将该非凸问题转化为可求解的双层凸优化问题;最后采用迭代优化算法对该问题进行求解.仿真结果表明：相较于原有算法,所提算法大幅减少了所需优化的参数数目,且在低信噪比时能实现约0.05 d B的性能提升.The ADMM（alternating direction method of multipliers） decoding algorithm based on improved penalty function can improve the performance of decoding algorithm,but it has too many parameters which need to be optimized.For this problem,an ADMM-LDPC decoding algorithm was proposed based on equilibrium constraint to solve it.At first,the 0-1 constraint of codes was converted into continuous equilibrium constraint,equivalently.After that,the original problem became a non-convex problem with equilibrium constraint. Then the problem was converted into double-layer convex problem which can be solved. Finally, the double-layer convex problem was solved by iteration optimization algorithm.The simulation results show that,compared with the original decoding algorithm,the decoding algorithm in this paper reduces the parameters which need to be optimized effectively and achieves about 0.05 d B performance improvement in low SNRs.

关 键 词：信息技术 低密度奇偶校验(LDPC)码 交替方向乘子法(ADMM) 均衡约束 ADMM惩罚译码 

分 类 号：TN915.04[电子电信—通信与信息系统]