| Spatial spectrum estimation is an important area in array signal processing ,which is developing rapidly in recent thirty years and emphasizes particularly on the capability of the spatial multi-sensor array system's the accurate estimation to the parameters of the interested spatial signals. To estimate the spatial parameters and the location of the signal sources is the spatial spectrum estimation's purpose, which is some areas'task, such as radar, communication and sonar etc. In theory, the technology can improve the accuracy of the estimation of the arrival angel and other parameters in the signal bandwith of the processing system. So it is widely used in radar, communication and sonar etc. In reality, due to the impact of the multi-path aspects, there are many coherent signal sources in space. But the general subspace direction of arrival (DOA) methods such as MUSIC and ESPRIT can not estimate the arrival accurate angel of the coherent signal sources because the signal subspace and the noise subspace influence mutually. Therefore, the DOA estimation of the coherent signal sources is a hot problem and it is a main point of the spatial spectrum estimation all the whole. Fluctuating targets such as planes always appear in the radar systems and ships fluctuate with the water wave. Thus the DOA estimation of the fluctuating targets is thought so much. The study of this paper includes two primary questions: the estimate of DOA methods and preprocessing methods of the coherent signals. We study the estimate of DOA methods and find that the existing methods must preprocess the coherence otherwise can not get the available estimation. In the other hand, the mature spatial-smoothing(SS) method can only be used in the unified linear array and the frequency-smoothing(FS) method have disadvantages of high complexity and huge computational cost for constructing the focus matrix. The excellent methods of coherent signals DOA estimation are even less. To solve these questions, here we propose a set of coherent signals DOA estimation methods based on fluctuating targets. The main tasks are generalized as: 1. Deeply study the DOA estimation methods and find that the DOA-matrix method can get more accurate estimation because it uses not only the eigenvalue information but also the eigenvector information, consequently the DOA-matrix method is a good latter-processing method. 2. Based on the analysis of fluctuating targets radar signals'characteristic and the existing coherence-preprocessing methods, a method which utilizes the short coherent temporal segment-smoothing to the fluctuating targets is proposed. Proved that the method can restrain the loss of array aperture and get the more accurate estimation. 3. Analyze the character of uniform-linear-array(ULA) ,L-uniform -linear-array(ULA) and uniform-circular -array(UCA), this paper proposes the coherent signal sources DOA estimation methods which are suitable to ULA,L-ULA and UCA. Compared with general approaches, the methods in this paper have many characters: 1. We select the short coherent temporal segments and increase the number of the coherent temporal segments to the fluctuating targets, so that the Doppler frequency in the same segment shifted slowly and can be seen no changes, otherwise the Doppler frequency-shifting occurs between the two segments. Thereby we can use statistical method to smooth the coherence but the manifold can not affect. 2. The foundation of the fluctuating targets'coherent signal model utilizes not only the signals'spatial information but also the signals'Doppler information. If we use MUSIC or ROOT-MUSIC method directly to estimation one dimension DOA, the result is not good. Think of the character of the L-ULA and UCA and the DOA-matrix method ,we propose the estimation method of two...
|
PDF Full Text Request