【作者】 王杰；

【导师】 孙志国； 李翼翰；

【作者基本信息】 哈尔滨工程大学 ， 电子信息（专业学位）， 2023， 硕士

【摘要】 在军事、医疗和地震预测等领域,波达方向(Direction of Arrival,DOA)估计有着广泛的应用。随着信息量的飞速增长,相比窄带信号,宽带信号能携带更多有用信息,其应用的场景越来越多,大多数在窄带信号下性能优良的DOA估计算法无法应用于宽带信号中,因此有必要提升宽带信号下DOA估计算法的性能。传统的DOA估计算法往往只能在高斯白噪声背景下有效,但在实际通信环境中存在的往往是色噪声并且在电子对抗以及多径等应用场景中,信源通常会由非相干源转变成相干信源,存在秩亏,无法得到正确的子空间,传统算法的估计效果不理想。因此,需要进一步提高色噪声下DOA估计的性能。综上,为进一步提高算法在色噪声背景下的性能,根据侦察端接收到的信号带宽大小,本文分别针对宽带相干信号和窄带信号的估计算法进行研究并对窄带信号进行测向系统的半实物验证,具体工作内容如下:(1)针对色噪声下基于协方差差分去噪的方法在宽带相干信号DOA估计中存在对相干信源数有限制的问题,提出一种基于噪声圆形特性去噪和Toeplitz矩阵重构的估计算法。首先,对接收到的信号求取协方差矩阵,在这其中利用噪声的圆形特性消除噪声,为能达到对协方差矩阵进行Toeplitz矩阵重构的要求,通过协方差矩阵相乘来构造新的数据协方差矩阵;其次,通过Toeplitz矩阵重构来解相干;最后,利用旋转信号子空间(Rotational Signal Subspace,RSS)算法准则构造聚焦矩阵,使用传播算子算法实现DOA估计。理论分析及仿真实验表明该算法的有效性,并且对相干信源数的奇偶没有依赖,同时该算法也适用于高斯白噪声下宽带相干信号DOA估计。(2)相比宽带信号,当侦察端接收到窄带信号时,为增加算法适用性并侧重于工程应用,针对色噪声下信号DOA估计精度低的问题,提出基于迭代去噪和信号子空间重构相结合的算法。首先,经过三次迭代去噪,消除部分噪声分量,其次对协方差矩阵进行特征值分解得到信号子空间,利用信号子空间重新构造新的观测矩阵,计算无噪数据协方差矩阵,最后,利用常规空间谱估计算法得到估计值。通过仿真验证,该算法的性能优于四阶累积量MUSIC(Fourth-order Cumulants Multiple Signal Classification,FOC-MUSIC)算法。并且提出一种新的测向系统半实物验证方法,利用信道模拟器、多通道宽带板卡等试验设备配合相应控制软件搭建测向系统平台,并将半实物证结果与理论仿真结果进行对比。更多还原

【Abstract】 DOA estimation has a wide range of applications in the military,medical,earthquake prediction and other fields.With the rapid growth of information,compared with narrowband signals,wideband signals can carry more useful information,and their applications are more and more scenarios,most of which cannot be estimated by DOA estimation algorithms with good performance under narrowband signals,so it is necessary to improve the performance of DOA estimation algorithms under wideband signals.Traditional DOA estimation algorithms are often effective only in the Gaussian white noise background,but in the actual communication environment there is often color noise and in electronic countermeasures and multipath and other application scenarios,the source is usually converted from a non-coherent source to a coherent source,there is a rank loss,can not get the correct subspace,the traditional algorithm is not ideal for estimation.Therefore,there is a need to further improve the performance and increase the applicability of wideband signal DOA estimation under color noise.In summary,in order to further improve the performance of the algorithm in the color noise background,the reconnaissance end,after receiving the signal,according to the bandwidth size.In this paper,the estimation algorithms for broadband coherent signals and narrowband signals are investigated and the semi-physical verification of the directional system for narrowband signals is carried out as follows:(1)To propose an estimation algorithm based on noise circular characteristic denoising and Toeplitz matrix reconstruction for the problem that the method based on covariance difference denoising under color noise has a limitation on the number of coherent sources in wideband coherent signal DOA estimation.First,the covariance matrix is obtained for the received signal,in which the circular property of the noise is used to eliminate the noise,and to be able to achieve the Toeplitz matrix reconstruction of the covariance matrix,the new data covariance matrix is constructed by multiplying the covariance matrix;second,the Toeplitz matrix reconstruction is used to solve the coherence;finally,and the rotational signal subspace(RSS)algorithm criterion is used to construct the focusing matrix,using The propagation operator algorithm is used to achieve DOA estimation.Theoretical analysis and simulation experiments show the effectiveness of the algorithm,and there is no dependence on the parity of the coherent source number,and the algorithm is also applicable to the DOA estimation of broadband coherent signals under Gaussian white noise(2)To increase the applicability of the algorithm and focus on engineering applications when narrowband signals are received at the reconnaissance end compared to broadband signals.An algorithm based on a combination of iterative denoising and signal subspace reconstruction is proposed to address the problem of low accuracy of signal DOA estimation under color noise.Firstly,after three iterations of denoising to remove the effect of noise,then the eigenvalue decomposition of the covariance matrix to obtain the signal subspace,use the signal subspace to reconstruct the new observation matrix,subtract the noise power matrix from the new covariance matrix to obtain the data covariance matrix,and finally use the conventional spatial spectral estimation algorithm to obtain the estimated value.The performance of the algorithm is verified by simulation to be better than the fourth-order Cumulants Multiple Signal Classification(FOC-MUSIC)algorithm.And propose a new method of semi-physical verification of the direction finding system,using the channel simulator,multi-channel broadband board and other test equipment with the corresponding control software to build the platform of the direction finding system,and compare the semiphysical proof results with the theoretical simulation results to verify the feasibility of the proposed method.更多还原