| Estimating the multi-parameter of array signals is an important area of signal processing. However, classical eigenstructure methods resort to estimating the array covariance matrix and performing the eigenvalue decomposition, which make the complexity high especially when the model order is large. Therefore, it is difficult to implement the real-time processing in the practical application. On the other hand, for the purpose of low complexity, it is often decompose a multi-dimensional problem into some 1-D problems, thus, coupling information must be employed to pair the respective members of each set of estimates. This dissertation is focused on the melioration of above two problems. The main work can be summarized as follows: The fundamental theory of the array signal processing, including the model of array system and the basic assumptions of array processing, is systematically addressed. Then, the analysis of the CRB of DOA estimation of signals with Spatial-Temporal model is presented. Furthermore, the idea of reduced-rank processing is given based on the theoretical analysis. The methods of multi-parameter estimation based on the beamspace processing are studied. Frist, two algorithms for the 2-D DOA estimation based on the two kind of conformal array and phase mode excitation are studied. After that, the 2-D DFT beamspace 2-D DOA estimation for URA is studied in detail, the performance of further reduced dimension processing is also discussed. Furthermore, the three dimensional frequency and 2-D DOA estimation in beamspace is studied, then, both the theoretical meaning and practical meaning of this algorithm are pointed out. The methods of multi-parameter estimation based on the propagator method (PM) are studied. The PM does not need the estimate of the covariance matrix and its eigendecomposition, therefore, it is more computationally efficient. First, a brief recall of PM is made. Then, four kind of multi-parameter estimation methods... |
PDF Full Text Request