散射信道中咬尾与非咬尾Turbo 码并行译码性能研究

Performance Analysis of Tail-Biting and Zero-Tail Turbo Codes with Parallel Decoding in Scattering Channels

摘要

散射通信由于存在较为严重的多径衰落，对通信设备的抗衰落能力提出更高要求，信道纠错编码作为无线通信中的重要部分，是提高散射信道通信传输性能的重要手段，Turbo码作为目前广泛使用的编码方式，它的终止方式将影响编码效率和译码实现的复杂度。本文针对散射信道环境，系统对比了咬尾和非咬尾Turbo码在不同码长、不同并行度下的并行译码性能，采用Max-Log-Map算法通过均匀分块与边界状态交换实现并行处理。仿真结果表明：在短码长（如N=128）下，咬尾码凭借无拖尾比特损失和均匀的译码增益，相比非咬尾码拥有约1dB的性能优势，且有效抑制了非咬尾码在高信噪比下的误码平层；随着码长增加，咬尾码的无边界优势逐渐凸显，在N=4096时相比非咬尾码获得约0.5dB的净增益，而在长码（N=8192）下两者性能趋于一致。此外，并行度对两种结构的影响显著不同：咬尾码性能对子段长度敏感，短码高并行度易导致收敛失效和平层出现，而非咬尾码对并行度变化表现出较强的鲁棒性，其性能主要受限于信道条件而非并行粒度。 该研究为散射信道系统中Turbo码的编码方式与译码架构的联合优化提供了参考依据。

Abstract

Severe multipath fading in scattering channels demands robust anti-fading capabilities from communication equipment, making error- correcting coding essential for maintaining transmission reliability. As Turbo codes are widely deployed, their termination strategy significantly dictates both coding efficiency and decoding complexity. This study evaluates the parallel decoding performance of Tail- Biting (TB) versus Non-Tail-Biting (NTB) Turbo codes across various code lengths and parallel degrees in scattering environments. We implement parallel processing using the Max-Log-Map algorithm with uniform block partitioning and boundary state exchange. Results show that for short codes (e.g., N=128), TB codes outperform NTB counterparts by approximately 1 dB, primarily due to the elimination of tail-bit overhead and uniform gain distribution, which also mitigates the high-SNR error floors typical of NTB codes. As code length increases, the boundary-free advantage of TB codes persists, yielding a 0.5 dB gain at N=4096. However, this gap vanishes at N=8192 as structural differences become negligible relative to the sequence length. Crucially, we observe distinct sensitivities to parallelism: TB performance degrades in short-code, high-parallelism scenarios due to insufficient sub- block convergence, whereas NTB codes remain robust, with performance largely dictated by channel conditions rather than parallel granularity. These findings offer practical guidelines for jointly optimizing termination methods and decoding architectures in scattering communication systems.

关键词

散射通信 / Turbo码 / 咬尾 / 非咬尾 / 边界状态交换

Key words

troposcatter communication / Turbo / zero-tail / tail-biting / boundary state exchange

参考文献

[1] 张涛, 刘莹, 孙柏昶, 等. 对流层散射通信及其应用[M]. 北京: 电子工业出版社, 2020.

[2] 李志勇, 秦建存, 梁进波. 对流层散射通信工程[M]. 北京: 电子工业出版社, 2017.

[3] 张功岑. 基于FPGA的LTE-A系统的Turbo编译码算法的并行化研究与实现[D]. 电子科技大学, 2014.

[4] 平磊, 面向5G通信的咬尾卷积码和Turbo码技术研究[D], 西安电子科技大学, 2017.

[5] H. T. Pai, “Sliding Window Algorithm for TBCC Decoding,” in IEEE Communications Letters, vol. 14, no. 12, pp. 1158-1160, December 2010.

[6] 3GPP. Evolved Universal Terrestrial Radio Access(E-UTRA);Multiplexing and channel coding:TS 36.212 V15.2.0[S]. 2018.

[7] 汪祖海, 短块长极低码率信道编码优化设计[D], 北京邮电大学, 2018.

[8] 赵旦峰, 李文意, 杨建华, 等. 分块并行 Turbo 码译码算法的研究[J]. 哈尔滨工程大学学报, 2004, 25(2): 209-212.

[9] 樊宏伦, 袁悦. 低码率Turbo最优编码设计[J]. 电子技术. 2025 (08):83-85.

[10] 董咏潮, 灵活全并行Turbo译码算法设计及实现[D]. 北京邮电大学, 2025.

[11] Martin-Vega, Francisco Javier, et al. “Further improvements in SOVA for high-throughput parallel turbo decoding.” IEEE Communications Letters 19.1(2014):6-9.