级联码迭代译码的定点量化设计

本文提出了一种新的结合对数压缩与均匀量化的算法,应用于级联码迭代译码中.通过分析原始输入信息和互信息的量化时的信息损失,将量化问题转化为考察基于信息损失比的多目标规划.分析压缩函数的参数,得到了量化设计的准则.同时本文阐述了迭代过程中计算转移概率使用结合源统计特性的查表法,指出互信息的动态量化设计可以基于不同迭代次数进行优化.仿真表明给出的量化算法能很好的跟踪浮点运算.

Design of fixed-point quantization in iterative decoder for concatenated code

A novel method combining logarithm compression with uniform quantization in iterative decoder of concatenated code is purposed.Based on the analysis of the quantized loss of the input information and the extrinsic information,it translates the quantization problem to multi-purpose programming grounded on the SDR(signal to distort ratio).The authors discuss the influence of the parameters of the compressed function and present design rules of quantization.Furthermore,a look-up table is used to calculate the transferring probability and dynamic design of extrinsic information is dependent on different iterative number.