非周期扩频信号多伪码盲估计技术研究

Blind estimation of pseudo-noise (PN) codes of spread spectrum signals is one of the key technologies to non-cooperative receivers. There have been many methods to estimate PN codes of periodic signals modulated only by short spreading code. This project studies blind estimation of multi-PN codes of non-periodic spread spectrum(NPSS) signals modulated by short spreading code and long scrambling code. NPSS signals received contain only a part of the PN codes and their structures are different from spectrum signals’, which spread by short PN code or by periodic long PN code, and the periodicity of PN codes could not be used to model the multi-channels of NPSS signals with multi-users to separate each user’ PN codes, so blind estimation of multi-PN codes for non-periodic spread spectrum signals is a hot also challenging subject in recent study. Triple correlation function(TCF) is insensitive to missing data, TCF peaks of an m-sequence have a close relation with its primitive polynomial and different m-sequence has different distribution of TCF peaks. So this project will study on the TCF characterizes of m-sequence and Gold sequence, detection methods of triple correlation peaks by applying goodness of fit testing, blind estimation methods of multi-PN codes for single user NPSS signals based on TCF, and blind estimation methods of multi-PN codes for multi-user NPSS signals based on blind signal separation and TCF. The completion of the project will help to improve the detection and estimation theory of spread spectrum signals and also have important theoretical significance and applicable value to the development of spread spectrum communication technology and communication countermeasure technology.

扩频信号的伪码盲估计是非合作接收机的关键技术之一，已有方法都只涉及仅扩频的周期扩频信号伪码盲估计。本项目研究扩频又加扰的非周期扩频（NPSS）信号多伪码盲估计。NPSS信号不包含完整周期伪码，信号结构不同于短码扩频信号和周期长码扩频信号，无法利用伪码周期性构造多用户NPSS信号的多通道模型分离各个用户伪码，所以NPSS信号多伪码盲估计是目前研究热点和难点。利用三阶相关函数（TCF）对数据缺失不敏感、m序列TCF峰与序列的本原多项式有密切联系和不同m序列TCF峰分布不同等特点，研究m序列和Gold序列的TCF峰特性、基于拟合优度检验理论的TCF峰检测、基于TCF共同峰的单用户NPSS信号多伪码盲估计、基于盲信号分离和TCF的多用户NPSS信号多伪码盲估计。本项目的完成有助于完善扩频信号检测与估计理论，对于扩频通信技术和通信对抗技术的发展具有重要的理论意义和应用价值。

本项目主要研究基于三阶相关函数的扩频信号多伪码盲估计和基于盲信号分离的扩频信号多伪码盲估计的理论与方法，进一步完善扩频信号检测与估计理论体系，为工程应用提供理论与技术支持。.理论上，首次证明了不同m序列三阶相关函数共同峰的存在性，分析得到了三阶相关函数共同峰的分布规律；证明了不同周期m序列的三阶相关函数共同峰坐标与其本原多项式之间的对应关系及正反向三阶相关函数共同峰之间关系等一系列性质。这些成果进一步完善了m序列三阶相关函数理论及特性，为研究扩频信号伪码盲估计提供了重要理论基础。.利用m序列的三阶相关函数特性，研究提出了直接序列扩频（Direct Sequence Spread Spectrum，DSSS）信号伪码估计算法四种。特别是，①利用不同周期m序列的三阶相关函数共同峰坐标与其本原多项式之间的对应关系，得到了周期/非周期短码扩频长码加扰的直扩信号（PLSC-DSSS/NPLSC-DSSS）的伪码估计算法；②首次将拟合优度检验理论应用于非周期长码直扩（NPLC-DSSS）和NPLSC-DSSS信号的伪码估计，进一步提高了伪码估计正确率。.利用三阶相关函数、独立分量分析（Independent Component Analysis，ICA）和矩阵理论，研究提出了直扩码分多址（Direct Sequence Code Division Multiple Access，DS-CDMA）信号多伪码估计算法六种。特别是，①基于张量分解的多天线周期长码直扩码分多址（PLC-DS-CDMA）信号的多伪码盲估计方法；②基于矩阵填充和三阶相关的周期短码扩频长码加扰的码分多址（PLSC-DS-CDMA）信号的多伪码盲估计算法；③利用Fast-ICA算法和三阶相关函数的非周期长码直扩码分多址（NPLC-DS-CDMA）/非周期短码扩频长码加扰的码分多址（NPLSC-DS-CDMA）信号的多伪码盲估计方法。.将深度学习应用于DS-CDMA信号伪码估计，研究提出了基于稀疏自编码器的PLC-DS-CDMA信号伪码估计。.本项目完成了研究目标和研究内容，在国内外学术期刊上发表论文13篇，其中SCI/EI论文7篇；在国际学术会议上发表论文1篇，被EI收录；获得授权发明专利4项，申请且目前已公开发明专利2项。