基于先进信号处理方法的通信信号调制识别技术研究
发布时间：2018-10-12 14:19
【摘要】：通信信号调制识别是指对接收信号自动处理并判定其调制类型的过程。作为信号检测与解调的中间环节，调制识别技术在认知无线电、智能解调器、电子侦察等各种民用及军事应用中扮演着重要角色。调制识别技术经过几十年的发展，虽然已经取得了很多成果，但随着工程化需求逐渐提高，无线通信信道环境日益复杂，仍有不少问题亟待解决。 本文重点致力于似然比调制识别算法的工程化应用和基于先进信号处理方法的调制识别技术的研究，论文的主要工作和创新性成果主要包括： 1.从通信信号处理的工程实用角度出发，针对似然比调制识别算法计算复杂度高的问题，提出了一种改进的快速算法。该算法引入硬件实现中预存查表的思想，通过一个查找表的地址读出待识别信号的似然函数值以便节省在线处理时间，为似然比调制识别算法的实时化应用提供了一种解决思路。试验结果表明算法能够在保证似然比算法最优性的基础上有效地节约时间成本，更能适应实时性要求高的应用场合。 2.在平坦衰落信道环境下，提出了基于一种自适应马尔可夫链蒙特卡罗（MCMC）技术——自适应Metropolis（AM）技术的调制识别算法。该算法能够在迭代过程中产生满足目标分布的未知参数和发送符号的各态历经样本，从而在实现似然函数的近似计算的同时完成参数估计，可实现调制识别与参数估计的一体化处理。相比于传统Metropolis-Hastings（MH）技术，AM技术避免了因建议分布函数选取不当造成的性能损失。仿真结果表明在平坦衰落信道环境下，基于AM技术的调制识别算法能够快速、精确的收敛，具有很好的识别性能。 3.将平坦衰落信道环境中基于AM技术的调制识别算法推广到多径信道环境中，为解决AM算法在未知参数维数较高时收敛速度放缓的问题，提出了基于单分量自适应Metropolis（SCAM）技术的调制识别算法。该算法在迭代过程中按顺序依次对未知参数向量中每个分量单独进行类似AM采样的操作。仿真结果表明在未知参数维数较高的情况下，，SCAM算法拥有更加优越的收敛性能，基于SCAM技术的调制识别算法在多径信道环境下识别性能良好。 4.在低信噪比环境下，提出了基于混沌理论的调制识别算法。文中对Duffing振子大尺度周期态特性进行了系统研究，考察了激励信号频率、幅度和相位对Duffing振子周期解的影响。根据大尺度周期状态下系统解随激励信号相位变化的规律，设计了基于Duffing振子的MPSK信号调制识别算法。并根据调制识别中的特征提取任务需求，对Duffing振子进行了模型优化，将优化后Duffing振子系统解的Poincaré映射作为分类特征，实现了低信噪比下的MPSK信号的调制识别。该算法仅仅利用载波频率的先验信息，无需进行码元同步，并且对信号幅度、载波初始相位以及载波频率的变化不敏感。仿真表明该算法具有较强的噪声免疫力，在信噪比较低的情况下仍能达到满意的识别效果。
[Abstract]:Communication signal modulation identification refers to a process of automatically processing a received signal and determining its modulation type. As the intermediate link of signal detection and demodulation, modulation recognition technology plays an important role in various civil and military applications such as cognitive radio, intelligent demodulator and electronic reconnaissance. After several decades of development of modulation recognition technology, although many achievements have been achieved, with the increasing engineering demand, the wireless communication channel environment is becoming more and more complex, and there are still many problems to be solved. In this paper, we focus on the engineering application of likelihood ratio modulation recognition algorithm and the research of modulation recognition technology based on advanced signal processing method. Including: 1. Engineering from communication signal processing Aiming at the problem of high computational complexity of likelihood ratio modulation recognition algorithm, an improvement is put forward. The algorithm introduces the idea of pre-stored look-up table in hardware implementation, reads the quasi-random function value of the signal to be identified through the address of a look-up table in order to save on-line processing time, and provides a real-time application of the likelihood ratio modulation recognition algorithm. The experimental results show that the algorithm can effectively save time cost on the basis of guaranteeing the optimality of likelihood ratio algorithm, and can meet the requirement of real-time performance. In the context of flat fading channel, a self-adaptive Markov chain Monte Carlo (MCMC) technique is proposed. The algorithm can generate the unknown parameters satisfying the target distribution and the state ergodic samples of the transmitted symbols in the iterative process, Compared with the traditional method of Metropolis-Hastings (MH), AM technology avoids the selection of the proposed distribution function. The simulation results show that the modulation recognition algorithm based on AM technology can rapidly and accurately converge in a flat fading channel environment. In order to solve the problem of slowing convergence speed when the unknown parameter dimension of AM algorithm is higher, a single-component adaptive Polis (SCA) is proposed in this paper. M) a modulation recognition algorithm of the technique, wherein each component in the unknown parameter vector is separated in sequence according to sequence in the iterative process The simulation results show that the SCAM algorithm has better convergence performance when the unknown parameter dimension is high, and the modulation recognition algorithm based on the SCAM technology is in multi-path. in a low signal-to-noise ratio environment, In this paper, a systematic study of Duffing oscillator's large-scale periodic state characteristics is carried out, and the frequency, amplitude and phase pairs of excitation signals are investigated. The effect of periodic solution of uffing oscillator. Based on the law of the phase change of the system solution with the excitation signal under the large-scale periodic state, Duffing oscillation is designed. According to the requirement of feature extraction task in the modulation recognition, the Duffing oscillator is optimized, the Poincare map of the optimized Duffing oscillator system is used as the classification feature, and the low signal noise is realized. the algorithm only uses a priori information of the carrier frequency, does not need to perform symbol synchronization, The simulation results show that the algorithm has strong noise immunity and is low in signal-to-noise ratio.
【学位授予单位】：西安电子科技大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN911.3
本文编号：2266446
[Abstract]:Communication signal modulation identification refers to a process of automatically processing a received signal and determining its modulation type. As the intermediate link of signal detection and demodulation, modulation recognition technology plays an important role in various civil and military applications such as cognitive radio, intelligent demodulator and electronic reconnaissance. After several decades of development of modulation recognition technology, although many achievements have been achieved, with the increasing engineering demand, the wireless communication channel environment is becoming more and more complex, and there are still many problems to be solved. In this paper, we focus on the engineering application of likelihood ratio modulation recognition algorithm and the research of modulation recognition technology based on advanced signal processing method. Including: 1. Engineering from communication signal processing Aiming at the problem of high computational complexity of likelihood ratio modulation recognition algorithm, an improvement is put forward. The algorithm introduces the idea of pre-stored look-up table in hardware implementation, reads the quasi-random function value of the signal to be identified through the address of a look-up table in order to save on-line processing time, and provides a real-time application of the likelihood ratio modulation recognition algorithm. The experimental results show that the algorithm can effectively save time cost on the basis of guaranteeing the optimality of likelihood ratio algorithm, and can meet the requirement of real-time performance. In the context of flat fading channel, a self-adaptive Markov chain Monte Carlo (MCMC) technique is proposed. The algorithm can generate the unknown parameters satisfying the target distribution and the state ergodic samples of the transmitted symbols in the iterative process, Compared with the traditional method of Metropolis-Hastings (MH), AM technology avoids the selection of the proposed distribution function. The simulation results show that the modulation recognition algorithm based on AM technology can rapidly and accurately converge in a flat fading channel environment. In order to solve the problem of slowing convergence speed when the unknown parameter dimension of AM algorithm is higher, a single-component adaptive Polis (SCA) is proposed in this paper. M) a modulation recognition algorithm of the technique, wherein each component in the unknown parameter vector is separated in sequence according to sequence in the iterative process The simulation results show that the SCAM algorithm has better convergence performance when the unknown parameter dimension is high, and the modulation recognition algorithm based on the SCAM technology is in multi-path. in a low signal-to-noise ratio environment, In this paper, a systematic study of Duffing oscillator's large-scale periodic state characteristics is carried out, and the frequency, amplitude and phase pairs of excitation signals are investigated. The effect of periodic solution of uffing oscillator. Based on the law of the phase change of the system solution with the excitation signal under the large-scale periodic state, Duffing oscillation is designed. According to the requirement of feature extraction task in the modulation recognition, the Duffing oscillator is optimized, the Poincare map of the optimized Duffing oscillator system is used as the classification feature, and the low signal noise is realized. the algorithm only uses a priori information of the carrier frequency, does not need to perform symbol synchronization, The simulation results show that the algorithm has strong noise immunity and is low in signal-to-noise ratio.
【学位授予单位】：西安电子科技大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN911.3
【参考文献】
相关期刊论文 前10条
1 吕铁军,郭双冰,肖先赐;基于模糊积分的通信信号调制识别方法研究[J];电子学报;2001年06期
2 王首勇,朱光喜,唐远炎;应用最优小波包变换的特征提取方法[J];电子学报;2003年07期
3 韩钢,李建东,李长乐,蔡雪莲;一种改进的基于累积量的MDPSK信号分类算法[J];电子学报;2004年10期
4 路鹏,李月;微弱正弦信号幅值混沌检测的一种改进方案[J];电子学报;2005年03期
5 陈卫东,杨绍全;基于循环累量不变量的MPSK信号调制识别算法[J];电子与信息学报;2003年03期
6 李杨,李国通,杨根庆;通信信号数字调制方式自动识别算法研究[J];电子与信息学报;2005年02期
7 柳征;王明阳;姜文利;周一宇;;一种新的贝叶斯调制分类算法[J];电子与信息学报;2006年07期
8 王彬;葛临东;徐立清;刘媛涛;;一种基于信道盲辨识和盲均衡的多径信道调制方式识别算法[J];电子与信息学报;2008年08期
9 谭晓衡;鄢海燕;苏萌;;基于自适应小波消噪的数字调制识别优化算法[J];电子与信息学报;2011年02期
10 阮秀凯;张志涌;;基于小波变换的调制自动识别新方法研究[J];南京邮电大学学报(自然科学版);2007年01期
相关博士学位论文 前1条
1 谢涛;外激励型混沌振子微弱信号检测方法研究[D];北京交通大学;2009年
本文编号：2266446
本文链接：https://www.wllwen.com/kejilunwen/wltx/2266446.html
