首页> 外国专利> Parallel decoding method for turbo decoding and turbo decoder using the same

Parallel decoding method for turbo decoding and turbo decoder using the same

机译:用于turbo解码的并行解码方法和使用其的turbo解码器

摘要

Parallel decoding method for turbo decoding according to the present invention, by dividing the input data signal into sub-blocks of a predetermined number and parallel decoding method for turbo decoding the respective sub-blocks that are parallel and independently, the predetermined number of times according to the performance specifications and by performing the iterative decoding process, a decoding process is repeated ; The odd-numbered sub-block performs the forward state metric calculation first, and the even-numbered sub-blocks comprising a reverse state metric calculation process, first; The odd-numbered sub-blocks comprising a reverse state metric calculation process, even-numbered sub-block performs the forward state metric calculations; And each sub-block is characterized by having a step for calculating the extrinsic information value. ; According to the present invention First, no time delay by eliminating the overhead of the training period when the parallel decoding for turbo decoding. Second, by using a boundary condition metric value calculated for inaccuracies in the initial state metric value that is a problem when the block division in the adjacent block is obtained the effect to put a longer training period than a conventional algorithm. Therefore, the performance degradation does not occur in a shorter length than the sub-block. Decoding delay is proportional to the length of the sub-blocks. Therefore, it is possible to implement turbo decoder with a smaller decoding delay, according to the present invention.
机译:根据本发明的用于turbo解码的并行解码方法,通过将输入数据信号划分为预定数量的子块,并通过并行解码方法,将根据预定次数的并行和独立的各个子块turbo解码。根据性能规格,通过执行迭代解码过程,重复解码过程;奇数子块首先进行前向状态量度计算,偶数子块首先进行反向状态量度计算过程。奇数子块包括反向状态度量计算过程,偶数子块进行前向状态度量计算;并且每个子块的特征在于具有用于计算外部信息值的步骤。 ;根据本发明,首先,通过消除用于turbo解码的并行解码时的训练周期的开销,没有时间延迟。其次,通过使用针对初始状态度量值中的不准确性计算的边界条件度量值,当获得相邻块中的块划分时,这是一个问题,其效果是比常规算法更长的训练周期。因此,不会以比子块更短的长度发生性能下降。解码延迟与子块的长度成正比。因此,根据本发明,可以实现具有较小解码延迟的turbo解码器。

著录项

  • 公开/公告号KR100627723B1

    专利类型

  • 公开/公告日2006-09-25

    原文格式PDF

  • 申请/专利权人

    申请/专利号KR20040047646

  • 发明设计人 신형식;김재석;정윤호;

    申请日2004-06-24

  • 分类号H03M7/40;

  • 国家 KR

  • 入库时间 2022-08-21 21:22:54

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号