Turbo高速编译码技术研究

Research on High-Speed Turbo Encoding and Decoding

本文提出了一种高速Turbo编译码方法。从算法改进和结构改进技术两方面进行研究,以期解决现有译码算法难以实现高速这一问题。在结构改进技术方面,采用分块思想,将分量编码器分成两块并行处理,速度提高一倍;在算法改进技术方面,一方面针对目前存在的复杂度较低、性能次优的Radix-4 Max-Log-MAP译码算法,通过尺度因子的补偿,得到了译码性能较好的SF-Max-Log-MAP算法。另一方面采用了HDA停止迭代准则,有效地减少了译码时延。

In this paper, one high speed Turbo encoding and decoding scheme is put forward. In order to achieve high-speed decoding, the improvements of algorithms and structure are discussed. In terms of the structure improvements, the parallel sub-block algorithm is adopted. Constituent Encoder is divided to two parts which are processed in parallel and now decoding twice the original speed. In terms of the algorithms improvements, the SF-Max-Log-MAP algorithm is proposed in order to enhance the decoding performance of Radix-4 Max-Log-MAP algorithna. On the other hand, the effective criterion HDA for stopping the iteration process is adopted which is to reduce decoding delay effectively.