Research On Detection And DOA Estimation Of Multiple Sources Based On Doublet Array

Array signal processing is used in several important application areas, such as military electronic countermeasures(ECM), radio communications, satellite navigation, seismic exploration, and medical imaging. For decades, scholars have carried out extensive research with various issues in array signal processing field and have achieved fundamental theories and basic methods. In practical applications, there are still numerous problems to be resolved. This dissertation offers an in-depth study for some of the issues, i.e., the structure of array, the received data model of array, multi-source detection technology and space parameter estimation, and obtain a lot of valuable work. Several methods are proposed and evaluated by both simulations and theoretical analysis. The main contributions are summarized as follows:1. The property of doublet array is firstly analysed and the data model of doublet array is present. The paper firstly propose the concept of the short-time subarrays array and short-time doublet array. The short-time doublet array provides the rationale basis that the received data of remote distribution non-coherent subarrays can be exploited together with each other. The property of doublet array is systematically studied and the advantages of doublet array in array signal processing is fully discussed. The estimation methods which are suitable for multiple subarray array are reviewed. Based on the current engineering implementation, the phenomenon of short-time coherence is analysed. Exploiting the phenomenon, the concept of short-time coherent subarrays array and short-time coherent doublet array are proposed, respectively. The short-time coherent doublet array provides the basic capability that the received data of subarrays can be further exploited.2. Detection technology for multiple sources based power is proposed. The shortcoming of the typical information theory criterion is analysed theoretically and the powerbased source enumeration methods in array signal processing are proposed. The present methods can avoid the shortcoming of information theory criterion. Based on the doublet array, the power of sources is estimated robustly. With the estimated power of sources, an excellent Power-AIC approach and average value determine method is suggested. Our work expands the applicable scope of time domain energy detection technology and presents a space power(energy) detection method. A post estimator for accuracy improvement is also proposed. The estimator can enhance the continuity of the steps of the source enumeration and space parameters estimation.3. The two-dimensional direction-of-arrival(DOA) estimation problem is studied. A2-D estimation algorithm with low computational complexity is proposed. The present algorithm employs the property of the centro-symmetry of the array and transforms the complex received data into real data, thus the computational load of the subsequent matrix operations can be reduced. By exploiting the eigenvalues and eigenvectors, the azimuth and elevation angles can be estimated. The one-to-one correspondence relationship between eigenvalues and eigenvectors enable azimuth and elevation angles to be auto pairing. Without estimation performance degradation, the proposed algorithm reduces the computational complexity with several aspects.4. The DOA estimation problem based on short-time doublet array is discussed. The received data model that far-field sources impinge upon the short-time doublet array is studied. The generalized property of short-time doublet array is analysed and emerged.By constructing the cost function, a spectral-search algorithm for far-field source DOA estimation is suggested. The present algorithm efficiently employs the data of each subarray which is considered to be noncoherent in the traditional sense, The algorithm can obtain better performance than the traditional approaches based on noncoherent subarrays array.5. Based on short-time doublet array, the estimation problem of parameters of arrival for near-field sources is studied and two estimation algorithms are proposed. The received data model for near-field sources based on short-time doublet array is analysed.The received data model based on short-time doublet array can achieve better accuracy with respect to traditional Fresnel model. For the single source case, a novel ESPRITlike algorithm with low computational complexity is proposed. By employing the symmetric distribution geometry of the short-time doublet array, a novel decoupled iterative algorithm is also presented to solve the problem of multiple sources case. The two algorithms exploit the received data of noncoherent subarrays array and obtain excellent performance. The two algorithms are suitable for single source real-time space parameters estimation and multiple space parameters estimation, respectively.