自适应信道化滤波与调制识别技术研究与实现

发布时间：2018-03-19 11:02

本文选题：软件无线电　切入点：数字下变频　出处：《解放军信息工程大学》2014年硕士论文　论文类型：学位论文

【摘要】：基于软件无线电架构的数字接收技术因其灵活性高、可复用性强等优点广泛应用于通信信号接收领域。本文重点针对软件无线电前端技术中的信道化滤波和后端技术中的自动调制识别及参数估计技术展开研究,并在信号处理平台上完成了相关算法和理论的系统实现。主要完成的工作和创新点包括以下几个方面:1.分析讨论了软件无线电接收技术中的基本理论,针对中频数字信号接收的预处理环节——数字下变频技术,以采样率变换三级结构为代表,进行了理论分析、基于FPGA的系统实现以及各级数据结果的谱分析,总结了实际接收机设计时应遵循的数字下变频多级结构设计准则。2.针对时变均匀分布的中频多路信号动态盲接收问题,首先提出了一种基于FPGA的动态信道化内核设计方案,其子信道数目、子信道阶数及原型滤波器系数均动态可配置,对设计原理、设计结构进行了分析介绍,通过Modelsim仿真及板上测试验证了其正确性。以此内核为基础,又提出了一种动态盲接收算法,可以在缺乏先验信息的条件下,对均匀分布的时变中频多路信号进行分信道接收,包含无盲区接收、频谱拼接、参数估计及校正接收等环节,对各环节的理论依据进行了分析推导,在课题信号处理平台上对该算法进行了仿真及实现。3.针对衰落信道下MFSK信号的符号速率估计问题,提出了一种基于小波脊线特征的符号速率估计算法,包含小波脊线提取、脊线滤波、过零点间隔聚类等环节,对各步的理论依据进行了数学推导。仿真实验表明,该算法能够有效克服多普勒频移、多径效应对符号速率估计的影响,获得了较高的估计精度。4.针对多径信道下通信信号的自动调制识别问题,首先在小波脊线的研究基础上,提出了一种MFSK和幅相调制信号的类间识别算法,相比传统算法,该算法在低信噪比多径信道下获得了较高的识别率,然后又提出了一种基于高阶累积量的幅相调制信号识别算法,该算法通过高阶累积量计算、非线性方程组构建以及迭代求解等途径,能够有效对信道干扰因子进行估计,相比目前的四阶矩商法,减小了信道因子分布对识别性能的影响,鲁棒性强。5.在课题信号处理平台上,完成了数字下变频、信道化滤波以及调制识别功能模块的构建,对整体系统进行了性能测试。测试结果表明,数字下变频和信道化滤波模块能够稳定的按照配置参数实现采样率变换、谱搬移及分信道滤波,调制识别模块正确完成了测试信号的调制方式鉴别,达到预期效果。
[Abstract]:Digital receiving technology based on software radio architecture has high flexibility, The advantages of strong reusability are widely used in the field of communication signal reception. This paper focuses on the channelization filtering in the front-end technology of software radio and the automatic modulation identification and parameter estimation in the back-end technology. And complete the system realization of related algorithms and theories on the signal processing platform. The main work and innovation include the following aspects: 1.The basic theory of software radio receiving technology is analyzed and discussed. In view of the pre-processing link of if digital signal reception-digital down-conversion technology, taking the three-level structure of sampling rate transformation as the representative, the system realization based on FPGA and the spectral analysis of the data result at all levels are carried out. In order to solve the problem of dynamic blind reception of if multi-channel signals with time-varying and uniform distribution, a dynamic channelized kernel design scheme based on FPGA is proposed. The number of subchannels, the order of subchannels and the coefficients of prototype filters are all dynamically configurable. The design principle and design structure are analyzed and introduced. The correctness of the design is verified by Modelsim simulation and on-board test. In this paper, a dynamic blind reception algorithm is proposed, which can receive time-varying if multi-channel signals with uniform distribution without prior information, including blind area free reception, spectrum splicing, parameter estimation and correction reception, etc. The theoretical basis of each link is analyzed and deduced, and the algorithm is simulated and realized on the platform of signal processing. 3. Aiming at the symbol rate estimation problem of MFSK signal in fading channel, A symbol rate estimation algorithm based on wavelet ridge feature is proposed, which includes wavelet ridge extraction, ridge filtering and clustering after 00:00. The theoretical basis of each step is derived. The algorithm can effectively overcome the influence of Doppler frequency shift and multipath effect on symbol rate estimation, and obtain a high estimation accuracy .4. aiming at the problem of automatic modulation recognition of communication signals in multipath channels, first of all, based on the research of wavelet ridges, In this paper, an inter-class recognition algorithm for MFSK and amplitude-phase modulation signals is proposed. Compared with the traditional algorithm, the algorithm achieves a higher recognition rate in low SNR multipath channels. Then, an amplitude and phase modulation signal recognition algorithm based on high order cumulants is proposed. This algorithm can effectively estimate the channel interference factor through the calculation of high order cumulants, the construction of nonlinear equations and iterative solution. Compared with the current fourth-order moment quotient method, the influence of channel factor distribution on the recognition performance is reduced. 5. On the signal processing platform, the digital downconversion, channelization filtering and modulation recognition function modules are constructed. The performance of the whole system is tested. The results show that the digital down-conversion and channelization module can realize the sampling rate conversion, spectral shift and sub-channel filtering stably according to the configuration parameters. The modulation recognition module correctly completes the modulation identification of the test signal and achieves the desired effect.
【学位授予单位】：解放军信息工程大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.3

【参考文献】