【作者】 谢锘；

【导师】 葛建华；

【作者基本信息】 西安电子科技大学 ， 通信与信息系统， 2010， 博士

【摘要】 现代无线电通信中电磁环境日趋复杂,同一频段内的信号越来越密集,大量同频或相干信号同时存在。现有的信号检测设备面临着越来越多的困难,例如,多个同频或相干信号的并行侦测、宽带信号的高分辨率测向、弱信号的检测、多目标的信号跟踪等问题,都是频谱检测系统中亟待解决的重大难题。目前,常用的经典空间谱估计算法已不能满足宽带相干信源DOA估计的需求,因此,研究宽带相干信源的精确测向技术具有重要的实际意义。本文对宽带相干信源的空间谱估计算法进行了深入研究,构建了宽带相干信源的数学模型,分析了宽带相干阵列信号的性质及其对经典空间谱估计算法的影响,分析了最大似然和相干信号子空间宽带空间谱估计算法等基本的宽带测向算法性质和特点。针对相干信号子空间类算法需要角度预估计问题,提出了一种基于角度扫描的宽带相干信号子空间算法,通过构造正交投影算子,保留了扫描角度处的信号,抑制其它方向的信号,新算法无需角度预估计。该算法较其他算法最突出的优点是,在宽带信源中心频率不一致的条件下仍能达到很好的测向效果。投影正交性测试算法(TOPS)在参考频点处信号子空间估计不准确时,空间谱估计结果会出现严重误差。针对这一问题,提出了解相干的改进TOPS空间谱估计算法(MTOPS),通过阵列协方差矩阵的共轭重构方法,实现了解相干,增强了算法的稳健性。高阶阵列累积量具有扩展阵列孔径和抑制噪声(包括高斯噪声和非高斯噪声)能力,在继承高阶累积量阵列信号处理算法优点的基础上,结合信号子空间聚焦算法提出了基于高阶累积量的宽带空间谱估计算法,该方法减小了噪声对空间谱估计的影响,有效提高了测向精度。针对ESPRIT算法要求阵列具有平移特性的问题,提出了基于高阶累积量的ESPRIT宽带测向算法,利用高阶累积量虚拟阵元,有效突破了ESPRIT算法对阵列形式的限制,成功地将其应用到均匀圆阵中。最后研究了宽带信号测向技术的工程实现,针对宽带信号处理特性对算法的实现进行分析,分别对系统组成、工作原理、硬件设计和技术指标进行详细论述,确定测向工作流程,并对功能的硬件实现进行合理划分。为减少运算时间,对空间谱测向算法中的数学运算进行了优化。同时给出了试验场测试结果,通过实际测向结果与理论仿真对比,验证了本测向系统的有效性。更多还原

【Abstract】 Electromagnetic environment is becoming more and more complex in modern wireless communication; the density of communication signals which locate in a common band is higher and higher. A lot of signals with the same frequency or coherent sources appear at one time. Detection equipments available face more and more troubles, for instance, detection of multiple signals which are coherent or have common frequency, high resolution direction finding of wideband sources, weak signal extraction, multi-target signals tracting, are all great difficult problems. At the present time, classical spatial spectrum estimation algorithm can not process wideband coherent sources. Consequently, it is meaningful to study precise direction finding techniques of broadband coherent sources.Broadband coherent sources spatial spectrum estimate algorithms are investigated extensively in the thesis. Mathematic model of wideband coherent source is built, by which we study characters of broadband coherent sources and its effect on spatial spectrum estimate algorithms.Considering coherent signal subspace algorithms have the drawback that it needs the pre-estimation of DOA. Coherent signal subspace algorithms which based on scanning angle are proposed. Through forming orthogonal projection operator, the information from scanning azimuths reserved, while information from other azimuths is suppressed. So the new method achieved good performance with no need of pre-estimation of DOA. Compared with other methods, the algorithm can achieve perfect performance when center frequencies of the wideband sources are not the same. When signal subspace estimation is not perfect in the reference frequency bin, the spatial spectrum estimation would be seriously distorted in the test of orthogonality of projected subspace (TOPS). To resolve the problem, the paper proposed a MTOPS algorithm, which enhances the algorithm’s robustness and provides the ability of decorrelation through conjugated reconstruction of the array covariance matrix.High order array cumulant has abilities of extending aperture of the array and suppressing noise (including Gaussian noise and non-Gaussian noise). Utilizing signal subspace focusing algorithm, we proposed wideband spatial spectrum estimation method based on high order cumulant, which can reduce the effect of noise on spatial spectrum estimation, the precision of direction finding is improved effectively.ESPRIT needs the shift-invariance property of the array, this thesis proposed a ESPRIT wideband direction finding algorithm based on high order cumulant. Using virtual element of array in high order cumulant, we break through the limitations of ESPRIT to array. The algorithm is applied to circular array which dose not have shift-invariance property successfully.Finally wideband signal direction finding engineering implementation techniques is studied in detail. Considering the characters of broadband signal processing, realization of the algorithms is analyzed. We discuss system composition, operation principle, hardware design, software design and technical specifications, determine working flow of direction finding, and form a reasonable division of hardware realization of functions. To reduce time of operation, the spatial spectrum estimation algorithm is optimized. Results of performance test in outdoor test site is given, comparison between actual direction finding results and that of simulations is drew, by which the efficiency of the direction system is shown.更多还原