基于非圆系数估计的宽线性波束形成算法研究

发布时间：2018-04-27 12:59

本文选题：阵列信号处理 + 宽线性波束形成　；参考：《中国科学技术大学》2014年博士论文

【摘要】：波束形成(Beamforming)技术是阵列信号处理领域中的一个重要研究方向,其利用空间多传感器所组成传感器阵列来对空间信号进行发射或接收,广泛应用于雷达,声呐,无线通信等领域。传统的波束形成技术假设信号是圆的、平稳的,仅仅利用了观测矢量的协方差矩阵,然而在通信领域大量的人工调制信号都是非圆的、非平稳的,不仅存在协方差矩阵而且存在伪协方差矩阵。基于圆信号假设下的波束形成技术虽然可以应用于非圆信号,但是其性能并不能达到最优。近些年现代信号处理领域中同时利用观测矢量及其共轭的宽线性(WL Widely Linear)技术的不断发展促使研究者们重新审视针对非圆信号的波束形成问题。宽线性波束形成技术通过分析信号的非圆性,同时利用了观测矢量的协方差矩阵和伪协方差矩阵,相应地一系列宽线性波束形成方法也陆续出现。本文在对宽线性波束形成技术进行概述的基础上,针对宽线性波束形成技术发展中存在的关键问题,开展基于非圆系数估计的宽线性波束形成算法研究。本研究工作的目的是在解决期望信号非圆系数估计问题的基础上,针对不同场景提出相应的宽线性波束形成算法。本论文研究工作的主要贡献与创新点归纳如下 1.针对最优宽线性波束形成算法中非圆系数难以估计的问题提出非圆系数估计算法,并得到了一个稳健的对角加载权重矢量。我们研究使用最小功率无畸变(MPDR)准则,将其转化为一个参数估计问题,通过求导得到初始估计后,分析干扰和噪声对初始估计的影响,最后得到一个修正的估计值。在得到非圆系数估计后,使用对角加载技术减小估计误差对算法性能的影响。理论与仿真实验同时证明所提算法对非圆系数有比较好的估计性能,而且能有效地应用于实际场景。 2.针对期望信号存在偏移频率的情况提出了频移宽线性波束形成算法。由于存在偏移频率的期望信号非圆系数随观测时间逐渐趋于零,共轭观测矢量与观测矢量中关于期望信号的相关信息是逐渐减少的。我们提出使用共轭循环观测矢量代替共轭观测矢量,其对应的共轭循环相关系数代替非圆系数进行宽线性波束形成。所提的算法在偏移频率为零时,等价于最优宽线性波束形成算法,在偏移频率不为零时,优于最优宽线性波束形成算法。此外针对期望信号是直线信号的情况,通过分析时间平均的非圆系数与偏移频率的关系,提出了两种估计偏移频率的算法。 3.当期望信号的偏移频率与干扰的偏移频率不同时,提出M阶频移宽线性波束形成算法。频移宽线性波束形成算法仅利用了期望信号的共轭循环相关函数,当期望信号的偏移频率与干扰的偏移频率不同时,考虑使用多个共轭循环相关观测矢量,通过增加M个约束使其干扰在对应于他们自己的循环频率上的输出为零。所提算法是频移宽线性波束形成算法的扩展版本,不增加约束的零阶算法对应着频移宽线性波束形成算法。当期望信号的偏移频率与干扰的偏移频率不同时,所提算法优于频移宽线性波束形成算法。 4.针对期望信号存在导向矢量失配的情况,引入非圆系数空间谱,提出了基于非圆系数空间谱估计的稳健宽线性波束形成算法。所提算法重构了干扰加噪声协方差矩阵和干扰加噪声伪协方差矩阵,剔除了扩展的协方差矩阵中的信号分量。在高信噪比和小快拍情况下,波束形成器输出信干噪比始终接近最优值。
[Abstract]:Beamforming (Beamforming) technology is an important research direction in the field of array signal processing. It is widely used in radar, sonar, wireless communication and so on. It is widely used in radar, sonar, wireless communication and so on. The traditional wave beam forming technique assumes that the signal is round, stable and only profit. The covariance matrix of the observed vector is used. However, in the field of communication, a large number of artificial modulation signals are noncircular and nonstationary. There is not only a covariance matrix but also a pseudo covariance matrix. The beamforming technology based on the circular signal hypothesis can be applied to the non circular signal, but its performance is not optimal. In the field of modern signal processing, the continuous development of the wide linear (WL Widely Linear) technology using observation vectors and its conjugate has prompted researchers to reexamine the beamforming problem for non circular signals. Wide linear beamforming technology uses the covariance matrix and pseudo covariance of the observation vector by analyzing the non roundness of the signal. Correspondingly, a series of wide linear beamforming methods have also appeared.
On the basis of the broad linear beamforming technology, this paper studies the wide linear beamforming algorithm based on the non circular coefficient estimation, aiming at the key problems in the development of the wide linear beamforming technology. The purpose of this study is to solve the problem of the estimation of the non circular coefficient of the expected signal and to the different scenes. The corresponding broad linear beamforming algorithm is proposed. The main contributions and innovations of this paper are summarized as follows.
1. the noncircular coefficient estimation algorithm is proposed for the problem of non circular coefficients in the optimal wide linear beamforming algorithm, and a robust diagonal loading weight vector is obtained. We use the minimum power distortion free (MPDR) criterion to transform it into a parameter estimation problem and analyze the interference after the initial estimation is obtained by the derivation. The effect of the noise on the initial estimation is finally obtained. After the estimation of the noncircular coefficient, the diagonal loading technique is used to reduce the effect of the estimation error on the performance of the algorithm. The theoretical and simulation experiments show that the proposed algorithm has a better performance for the non circular coefficients and can be effectively applied to the actual scene.
2. a frequency shift width linear beamforming algorithm is proposed for the frequency shift frequency of the expected signal. The correlation information between the conjugate observation vector and the observed vector is gradually reduced because the expected signal non circle coefficient of the expected signal tends to zero with the observation time. We propose to use the conjugate cycle observation vector. The proposed algorithm is equivalent to the optimal wide linear beamforming algorithm when the offset frequency is zero. The proposed algorithm is superior to the optimal wide linear wave beam formation algorithm when the offset frequency is not zero. In addition, the desired signal is a linear signal. By analyzing the relationship between the time averaged noncircular coefficient and the offset frequency, two algorithms for estimating the offset frequency are proposed.
3. the M order frequency shift width linear beamforming algorithm is proposed when the expected signal's offset frequency is different from the interference frequency. The frequency shift width linear beamforming algorithm only uses the conjugate cycle correlation function of the expected signal. When the expected signal's offset frequency is different from the interference frequency, the multiple conjugate cycle correlation concept is considered. The proposed algorithm is an extended version of the frequency shift width linear beamforming algorithm, and the zero order algorithm of the constraints corresponds to the frequency shift width linear beamforming algorithm. The offset frequency of the expected signal number is different from the interference frequency. The proposed algorithm is better than the frequency shift linear beamforming algorithm.
4. a robust wide linear beamforming algorithm based on the noncircular coefficient spatial spectrum estimation is proposed in the case of a non circular coefficient space spectrum. The proposed algorithm reconstructs the interference plus noise covariance matrix and the interference plus noise pseudcovariance matrix, and eliminates the signal points in the extended covariance matrix. In the case of high SNR and small snapshot, the output signal to noise ratio of the beamformer is always close to the optimal value.

【学位授予单位】：中国科学技术大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN911.7

【参考文献】