一种节省存储资源的Turbo码译码算法

针对Turbo码译码过程需要占用大量存储资源的问题，在Max—Log-MAP算法的基础上介绍了一种节省存储资源的译码方法^1]。通过逆运算在反向递推计算后向路径度量的过程中同时计算部分前向路径度量，节省了近50％前向路径度量的存储资源；通过使用一种性能良好的线性运算代替查表法求得Jacobian（雅可比）对数函数，避免了查找表资源的使用。在避免路径度量溢出进行的归一化操作中，采用了一种有效的方案，进一步节省了计算资源。实验表明，文章方法在有效节省存储资源的同时保证了良好的译码性能。

A Decoding Turbo Code Algorithm of Saving Memory Resources

Aiming at a great deal of memory resources used in decoding Turbo Code, a new decoding algorithm is proposed based on Max - Log - MAP algorithm. Nearly 50% memory resources for forward path measurement which are calculated with backward path measurement through backsetp anticalculate are saved. To avoid look - up resource, a linear arithmetic is used for calculating Jacobian logarithm function. Moreover an effective method is applied to save calculational resource in unitary path measurement. The experiment shows that the proposed algorithm not only saves memory resources, but also has good performance on decoding.