| The signal source localization is one of the most important research aspects in array signal processing, whose purpose is to deal with the data received from the antenna array, reduces the interfering signal and noise signal, and estimates the spatial location information of the interested source. According to the distance between the source and the receiver array, the source localization can be classified into far-field source localization and near-field source localization.Since the 1980s, the source localization technology research is focused on the far field source; by contrast, the research of parameter estimation theory for near-field was relatively less. In recent years, however, the theoretical study of near-field source parameter estimation, which is a new research focusing on the array signal processing, has gradually attracted the attention of scholars. After a development over one decade, many scholars have studied the parameters estimation problems of near-field and have achieved many excellent achievements. Nevertheless, there still remain many problems to be solved in some aspect, such as parameter matching, computation, estimation accuracy, coherent sources and broadband non-stationary signals etc. In this dissertation, the parameters estimation of near-field source is studied from the following six aspects respectively, parameters pairing, array aperture,2-D polarization sensitive array, non-Gaussian noise computation, and broadband non-stationary signals, by means of second order statistics, fractional lower order statistics, fractional Fourier transform and higher order cumulant.The creative work of this dissertation focuses on five aspects as follows.1. Utilizing the four order cumulant and non-centro-symmetric cross array, proposes bearing, elevation and range parameters estimation algorithm of near-field based on whiten dimension reduction, subspace decomposition and matrix pencil. The algorithms adoptes the non-centro-symmetric array construction, doesn't require any spectral peak search and additional parameter matching, can directly obtain the closed form solution of the parameters to be estimated, and has the features of high estimation accuracy, low complexity, and high utilization of the array aperture.2. Utilizing the advantages of anti-interference ability and higher resolution for the polarization sensitive array, by means of second-order statistics and fourth-order cumulant as the main tools of mathematical analysis, studies the parameter estimation problem of near-field based on 2-D polarization sensitive array, proposed one algorithm to estimate elevation, bearing, range, polarization information of near-field based on second order statistics and three algorithms based on four-order cumulant, and solved some problems effectively, such as high-efficiency array aperture, low computation and low computational complexity.3. Aiming at the impulsive noise features for no second-order statistics and no higher order cumulant, studies the 2-D parameters estimation problem of near-field in SaS Stable distribution impulsive noise, proposes a FLOS-ESPRIT algorithm based on fractional lower order statistics. The algorithm can inhibit the effect of impulsive noise, and has good effect in additive white Gaussian noise too.4. For the problem of large computation, proposes two 2-D parameter estimation algorithm of near-filed:PM-MUSIC and PM-ROOT-MUSIC. The algorithms used the noise subspace estimated by propagator method instead of orthogonal solution of SVD, and solved the contradictions between estimation accuracy and computation of MUSIC algorithm.5. For two-dimensional parameters estimation problem of the wideband near field sources of non-stationary, proposes a bearing and range parameters joint estimation algorithm of near field wideband non-stationary sources based on FRFT-MUSIC. Its characteristics are no cross-term interference, have high estimation accuracy, and can be applied for additive noise environment.This dissertation consists of eight chapters.1. It introduces the development of array signal processing technology, summarizes the theories of near-field parameter estimation as well as the current research situation at home and abroad, proposes the content to be studied and solved, and defines the research content of this dissertation.2. It introduces the fundamental knowledge of matrix theory and the polarization sensitive array, introduces the definition and nature of fractional lower order statistics, fractional Fourier transform and higher-order cumulant.3. It studies the localization algorithm in scalar sensor array based on four-order cumulant. Aiming at the problems of estimation accuracy and complexity, it proposes four parameters estimation algorithm without parameter pairing:(1) A joint estimation algorithm of bearing and range for near-field based on four-order cumulant and subspace decomposition, the algorithm adopts non-centro-symmetric array, and the aperture loss is effectively avoided, the parameters can be paired automatically, reduces the complexity of computation; (2) Above algorithm extends to the 2-D receiver array, proposes 3-D parameter joint estimation algorithm without parameters pairing for near-field sources based on non-centro-symmetric cross array; (3) A 3-D parameter joint estimation algorithm of near-field sources based on MP, the proposed algorithm adopts non-centro-symmetric cross array, so the array aperture loss is effectively avoided. Furthermore, the algorithm estimates the parameters of near field by generalized eigenvalue decompose; the parameter pairing information is contained in the magnitude value of the generalized eigenvalue, so doesn't require the extra parameter pairing method of various parameters, and reduces the complexity of computation. (4) Proposed bearing, elevation and range parameters estimation algorithm of near-field based on whiten dimension reduction, the algorithm need not construct high dimensional cumulant matrix, parameters pairing automatically, reduces the computation and complexity, and improves the utilization of the array aperture.4. It studies the localization algorithm in 2-D polarization sensitive array based on second order statistics and four-order cumulant:(1) Proposed bearing, elevation, range, and polarization information joint estimation algorithm of near-field based on second order statistics, its features are small computation, high estimation accuracy, and requires a simple parameters pairing only. (2) Proposed 5-D parameters joint estimation algorithm of near-field based on four order cumulant, it need not construct high dimensional cumulant matrix, pairing parameters easily and low complexity. (3) Proposed a joint 5-D parameters estimation algorithm of near-field sources based on four-order cumulant and subspace decomposition, adopts non-centro-symmetric cross polarization sensitive array, avoids the aperture loss effectively, without parameters pairing, reduced the complexity. (4) Proposed 5-D parameters estimation algorithm of near-field based on four order cumulant and joint diagonalization technology, adopts non-centro-symmetric cross polarization sensitive array, need not construct high dimension cumulant matrixes, estimates the signal parameters by whiten dimension reduction and joint diagonalization technology, without parameters pairing, reduces the computation and complexity, improves the utilization of the array aperture.5. It studies the localization algorithm of near-field in impulsive noise environment, proposes a FLOS-ESPRIT algorithm based on fractional lower order statistics to estimate bearing and range parameters of near-field, The algorithm adopts the center symmetric uniform linear array, can inhibit the effect of impulsive noise, and apply in additive white Gaussian noise environment too.6. It studies the propagator method, and proposes two algorithms of 2-D parameters estimation for near-field:PM-MUSIC and PM-ROOT-MUSIC, the algorithms use the noise subspace estimated by propagator method instead of orthogonal solution of singular value decomposition, and solve the contradictions between estimation accuracy and computation of MUSIC algorithm for reduce computational purposes.7. It studies the parameter estimation problem of broadband non-stationary of near field source, proposes a bearing and range joint estimation algorithm of near field source based on FRFT-MUSIC, and it characterizes of high estimation accuracy in low SNR.8. A brief summary of the dissertation is given, prospects the future work, and proposes the problems to be studied in future.
|
PDF Full Text Request