改善3D-Turbo码错误平台性能的不规则抽取方案设计研究
Irregular Pattern Design to Improve Error Floor for 3D-Turbo Codes

针对3D-Turbo码中仍然存在的低码重码字，以全零迭代译码算法作为码字距离谱分析工具，识别低码重信息比特和校验比特中"1"的分布，优化设计不规则抽取模式，进一步提高最小汉明距离，进而改善错误平台性能。最后，以一种3D-Turbo码为实例，利用文中所提的改进抽取方式，将最小码重从18提高到20，证明了改进抽取方案的有效性。
In 3D-Turbo codes, part of parity bits from conventional Turbo encoder are encoded again by a post-encoder to reduce the number of low-weight codes and achieve better error floor performance. Considering the residual low-weight codes in 3D-Turbo codes, an irregular pattern design is proposed to further improve the error floor. All-zero iterative decoding algorithm is firstly utilized to identify the distribution of "1" in low-weight codes. Then an irregular pattern is optimized based on the distribution to eliminate low-weight codes. Finally, taking a practical 3D-Turbo codes, we adopt the proposed irregular pattern to increase the minimum Hamming distance from 18 to 20, which validates the effectiveness of our contribution