MPSK信号调制方式识别与参数估计

发布时间：2018-02-21 13:05

本文关键词： MPSK 调制识别参数估计盲估计　出处：《北京邮电大学》2015年硕士论文　论文类型：学位论文

【摘要】：一直以来,通信信号调制方式的识别和参数估计都是通信领域的研究热点。信号的识别与估计在军用、民用领域具有重要意义,比如在电子对抗、情报侦察、无线电资源管理、防灾减灾和地质勘探等方面都有广泛应用。数字信号具有抗干扰、易于加密、安全性高等优点,如今已经广泛应用,发展出了许多种数字调制方式,因此对数字调制信号的估计更有意义。相位调制信号(MPSK)作为一种重要的数字调制信号,应用也十分广泛,本文就几种常用的相位调制信号展开研究,针对这类信号的参数盲估计和盲识别进行了较为深入的探讨。本文开篇交代了信号参数估计和调制识别的研究背景和意义,简要梳理了针对该课题研究的发展过程,介绍了相关的理论知识和算法。论文随后阐述了MPSK信号的调制产生过程,具体分析了BPSK.QPSK、 OQPSK、UQPSK和8PSK信号的基本特征,并借助星座图对这几种信号做了对比讨论。第三章详细讨论了MPSK信号的参数估计,包括带宽、载频、符号速率和信噪比。对于带宽估计,采用基于Welch变换的功率谱估计法,估计误差可达1%以下,高信噪比情况下误差能够达到3%。左右。对于载频估计,采用M次方谱估计法,当信噪比大于7.5dB时,估计误差可达10-5以下,最小误差可达3×10-6。符号速率估计采用延迟相乘法,在较大信噪比范围内能够得到准确估计值,且对噪声不敏感。对于信噪比的估计,分析了功率谱估计法、二阶矩四阶矩估计法,对二者的局限性做了说明。针对自相关矩阵奇异值分解估计法进行了详细阐述,指出了该算法的适用条件,并对不符合条件的带通滤波信号提出了一种改造方案。该方案能够有效估计接收端经过带通滤波之后的信号信噪比,信噪比低于20dB时估计误差在1dB以内,当信噪比高于20dB,低于30dB时误差在2dB以内。论文最后研究了MPSK信号调制方式的识别方法。针对相位调制信号的特点,介绍了谱线特征法、星座图法、相位直方图法等算法的特点和局限性。阐述了高阶统计量的基础知识,由于高斯白噪声的高阶累积量均为零,所以高阶累积量算法理论上能够消除白噪声对信号识别的影响。以此为基础构造了用于MPSK信号识别的特征参数,设计了信号识别流程,分别针对BPSK与UQPSK、QPSK与8PSK、QPSK与OQPSK三组信号做了对比仿真,仿真结果表明能够达到较好的识别效果：取符号长度为1024,当信噪比大于15dB时,BPSK和UQPSK信号的识别准确率可达95%以上；当信噪比大于10dB时,QPSK和8PSK信号的识别准确率可达95%以上；当信噪比大于5dB时,QPSK和OQPSK信号的识别准确率可达90%以上,当信噪比高于15dB时,识别准确率可达100%。
[Abstract]:All along, the identification and parameter estimation of the modulation mode of the communication signal are the research hotspot in the field of communication. The identification and estimation of the signal is of great significance in the military and civil fields, such as electronic countermeasures, intelligence reconnaissance, radio resource management, etc. The digital signal has many advantages such as anti-jamming, easy encryption, high security and so on. Now it has been widely used and many kinds of digital modulation methods have been developed. Therefore, the estimation of digital modulation signal is more meaningful. As an important kind of digital modulation signal, the phase modulation signal MPSKK is also widely used. In this paper, several commonly used phase modulated signals are studied. The blind parameter estimation and blind recognition of this kind of signal are discussed in detail. At the beginning of this paper, the research background and significance of signal parameter estimation and modulation recognition are explained, the development process of the research on this subject is briefly reviewed, and relevant theoretical knowledge and algorithms are introduced. After that, the modulation process of MPSK signal is described, and the basic characteristics of BPSK.QPSK, OQPSK UQPSK and 8PSK signal are analyzed in detail, and these signals are compared and discussed with the help of constellation diagram. In chapter 3, the parameter estimation of MPSK signal is discussed in detail, including bandwidth, carrier frequency, symbol rate and signal-to-noise ratio. For bandwidth estimation, the power spectrum estimation method based on Welch transform is used, and the estimation error can be less than 1%. In the case of high signal-to-noise ratio, the error can reach about 3. For carrier frequency estimation, M power spectrum estimation method is used. When the SNR is greater than 7.5 dB, the estimation error can reach 10 ~ (-5) and the minimum error can reach 3 脳 10 ~ (-6). In the range of signal-to-noise ratio (SNR), the accurate estimation value can be obtained, and it is not sensitive to noise. For the SNR estimation, the power spectrum estimation method and the second-order moment fourth-order moment estimation method are analyzed. In this paper, the limitations of the two methods are explained. The method of singular value decomposition of autocorrelation matrix is described in detail, and the applicable conditions of the algorithm are pointed out. A modified scheme is proposed for the nonconforming band-pass filter signal, which can effectively estimate the signal-to-noise ratio of the receiver after band-pass filtering, and the estimation error is less than 1dB when the SNR is less than 20dB. When the SNR is higher than 20 dB and less than 30 dB, the error is less than 2 dB. Finally, the recognition method of modulation mode of MPSK signal is studied. According to the characteristics of phase modulated signal, the spectral line characteristic method and constellation method are introduced. The characteristics and limitations of the algorithms such as phase histogram method are discussed. The basic knowledge of high-order statistics is expounded, because the high-order cumulants of Gao Si white noise are all zero. Therefore, the high-order cumulant algorithm can eliminate the influence of white noise on signal recognition theoretically. Based on this, the characteristic parameters for MPSK signal recognition are constructed, and the signal recognition flow is designed. Three sets of signals of BPSK and UQPSK QPSK and 8PSK QPSK and OQPSK are compared and simulated respectively. The simulation results show that the recognition accuracy of BPSK and UQPSK signals can reach more than 95% when the symbol length is 1024 and the signal-to-noise ratio is more than 15dB. The recognition accuracy of QPSK and 8PSK signals can reach more than 95% when SNR is more than 10dB, 90% when SNR is greater than 5dB, and 100dB when SNR is higher than 15dB.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN911.3

【参考文献】