【作者】 刘世刚；

【导师】 葛临东；

【作者基本信息】 解放军信息工程大学 ， 信号与信息处理， 2008， 博士

【摘要】 近十几年来,随着短波新体制新技术的日益发展,突发通信在短波通信领域,尤其在短波军事通信领域获得了越来越多的应用。然而对于该类信号的盲(非协作)接收技术,特别是短突发信号的盲接收技术的研究相对滞后,另外,计算复杂度低、有利于工程实现的算法较少。本文围绕这一应用背景,结合实验室承担的军队大型重点工程研究项目,针对短波突发信号的盲接收所涉及到的关键技术,从算法研究和工程实现两个方面进行深入地探讨。突发信号的存在性检测是短波突发信号盲接收首先要解决的问题。本文从工程实现的角度出发,将信号的存在性检测分为判决统计量的选取、自适应门限的调整和判决算法的确定三个相对独立又相互影响的部分,并指出在实际应用中减小虚警概率和漏警概率更有效的方法是设计合适的自适应门限算法和判决算法。同时针对短波突发信号的特点,分别给出了一种较为实用的自适应门限算法和判决算法。符号盲同步是短波突发信号盲接收的关键技术之一。本文总结了目前用于突发信号的非数据辅助前向符号定时估计算法,分析和仿真了各种算法的性能,指出了在实际应用中应注意的问题。在此基础上,提出了一种新的基于每符号两个样点的非数据辅助的前向符号定时估计算法。新算法研究了以两倍符号速率采样的基带信号经过非线性变换后在频域混叠的情况,然后,在非线性变换之前,直接采用抗混叠滤波器,将可能在非线性变换之后出现混叠的那部分信号滤掉。这样,以两倍符号速率采样的基带信号,经过抗混叠滤波后,就可以直接利用传统的平方律非线性算法(当然,也可以采用一次幂非线性法或对数非线性法等)估计定时相位偏差。新算法对于信号及应用环境的适应性更强。在短波数字信号盲接收系统中,盲均衡器的设计是至关重要而又相对困难的一个部分。尤其是对于短波突发信号设计盲均衡器,难度还是相当大的。其关键问题在于如何能在几百、甚至几十个符号内使盲均衡算法收敛。本文归纳和总结了目前盲均衡领域的研究现状,从实际应用的角度出发,重点分析了较为实用的Bussgang类盲均衡算法,并指出对于突发信号、尤其是短突发信号的盲均衡,常规Bussgang类盲均衡算法都很难在一个突发内完成收敛。为了解决快速收敛问题,经认真分析将研究重心放在了基于数据重用思想的Bussgang类盲均衡算法上,该领域的研究刚刚起步,成熟的算法相对较少。本文根据数据重用方法的不同,将该类算法分为三种,并重点研究了其中的两种:基于多数据向量的重叠重用和基于多数据向量的循环重用。对于前者,分别给出了基于仿射投影的CMA盲均衡算法和基于集员滤波的CMA盲均衡算法,加快了CMA的收敛速率;而对于后者,其算法结构更利于实时实现,并且本文以大量的仿真实验及分析证实了该类算法的有效性,其在一定的信道条件下完全可以在几十、甚至十几个符号内达到收敛,非常适合短突发信号的盲均衡。载波同步在数字信号的相干接收中是不可或缺的技术环节。而对于突发信号的盲接收来讲则必须采用非数据辅助的前向载波同步技术,主要包括频偏估计和相偏估计。另外,考虑到短波数字信号的实际接收条件,还应关注算法的抗噪声能力。本文归纳总结了目前常用的非数据辅助前向频偏和相偏估计算法,分析和仿真了各种算法的性能,在此基础上,提出了一种新的载波频偏估计算法。新算法首先利用相邻码元判决点处的相位差粗略估计载波频偏,从而获得宽的捕获范围;然后,再利用长时延自相关函数精细估计载波频偏,以便获得高的估计精度。新算法以相对较低的计算复杂度在一定程度上解决了相位折叠和低信噪比下的频偏估计问题。同时为了更好地解决相偏估计中由噪声引起的相位跳变问题,针对短波信号分别提出了一种改进算法和一种全新算法,两种算法都可以有效地降低相位跳变的概率,提高算法的抗噪声性能。最后,结合实验室承担的科研任务,针对短波窄带信号的实时盲接收,研制了两种基于软件无线电的短波窄带多通道数字信号处理平台,并将本文所研究的盲接收算法应用于这两种数字信号处理平台,实现短波窄带信号的实时盲接收。更多还原

【Abstract】 In recent years burst communications, due to its obvious advantages, have become more and more popular in modern shortwave communications, especially in military communications. Whereas the corresponding technologies of blind (non-cooperative) receiving for burst signals, especially for very short burst signals, are still not so mature as that for conventional signals. Just on above background, this paper is focusing on a study and implementation of the key technologies of shortwave burst signal blind receiving while laying emphasis on the development of high-efficient algorithms with lower computational complexity. The work finished in this paper is a part of a large scale army project of research undertaken by the lab the author works with.The existence detection of burst signal, which is the first thing to begin with, can be considered to consist of decision statistic selection, adaptive shrehold setting and decision algorithm development. Two practical algorithms for threshold setting and decision making respectively are proposed which can considerably reduce the false-alarm probability and false-dismissal probability by appropriate design of the threshold and the corresponding decision making mechanism.Blind symbol synchronization is one of key technologies of the blind receiving of the shortwave burst signals. Severl important NDA(nondata-aided) feedforward symbol timing estimation algorithms are analyzed and compared with each other in terms of their performances. Based on above discussion a novel NDA feedforward symbol timing estimator employing two samples per symbol is proposed. The new estimator can eliminate the spectrum overlap by the use of anti-aliasing filter, which can make the conventional non-linear algorithms, such as squared processing etc., safely be used in cases of two samples per symbol. Simulations have proved the performance improvement of the proposed estimator over the former one.Blind equalization is another one of key technologies of the blind receiving of the shortwave burst signals, and successful design of the blind equalizer is also a challenging issue in the field, especially for very short burst with only several hundreds or even decades of symbols. A variety of typical blind equalization algorithms are discussed and summarized with the emphasis laid on the Bussgang blind equalization algorithms. And then, the Bussgang blind equalization algorithms based on data-reusing are analyzed, which can generally be classified into three categories, and two of them are discussed in depth. A novel CM algorithm based on affine projection and data-reusing is proposed which, compared with CMA without data reusing, can converge much faster, however the steady-state error is increased. To solve this problem another novel CM Algorithm based on set-membership filtering and data-reusing is proposed. Besides, a blind equalization algorithm based on repeated data-reusing is proposed, which is of lower computational complexity. Simulation results have shown that the algorithm can converge even within dozens of symbols, thus very suitable for shortwave burst signals.Carrier synchronization is also critical and indispensable for blind receiving. Conventional algorithms for estimations of carrier and phase offset are summarized, simulated and compared to each other. Based on the discussion a new algorithm with wide estimation range, high accuracy and low-complexity for carrier offset estimation is proposed. Firstly, the capturing-range is widened by using the phase difference between consecutive symbols, and then higher estimation accuracy is obtained based on correlation function with longer delay while the phase folding and low SNR problems are overcomed. To solve the equivocal phase problem in phase offset estimation caused by residual frequency offset and noise, a modified and a novel algorithms are proposed, which can greatly reduce the phase hopping probability and improve the anti-noise performance.Finally, two blind receiving platforms based on Software Defined Radio (SDR) for shortwave narrowband multi-channel signals are developed. All of the blind receiving algorithms proposed in this dissertation are applied to the two platforms, and test results have proved the validity of the whole systems.更多还原