外信息在Turbo迭代译码中的使用研究

Research on Using Extrinsic Information in Iterative Decoding of Turbo Codes

以易于工程实现的Log-MAP算法为基础,深入研究了两种不同的外信息使用方式:Berrou方案及Robertson方案,指出两种外信息使用方式具有统一性,不同的是迭代译码过程中对外信息使用权重的处理;在相同约束条件下对基于两种外信息使用方式的Turbo码性能进行蒙特卡罗仿真,结果表明Robertson方案从整体上要略优于Berrou方案。

Based on Log-MAP algorithm of easy engineering implementation, the two different methods of using extrinsic information, the Berrou and Robertson schemes, are studied deeply in the paper. The study points out that the two methods are consistent, and the only difference lies in the treatment to the scale of weight applied to the extrinsic information in iterative decoding. Under the same constraint condition, by Monte Carlo means the performances of Turbo codes with extrinsic information in different fashions are simulated, and the result shows that Robertson method is better than Berrou method from the overview.