Research Of Technology About DOA Estimation For Super-multiple Wideband Sources

Super-resolution DOA (Direction-Of-Arrival) estimation is a hot topic in the spatial spectrum estimation field. The reasons why it has received considerable attention are the superior estimation performance and the wide range of applications. Most of super-resolution DOA estimation techniques were proposed under ideal assumptions. In real environments, however, due to the presence of super-multiple and (or) wideband sources, most of them have an unsatisfying show on precise estimation or even turn to be invalid. In this thesis, some problems about the super-resolution DOA estimation of super-multiple wideband sources have been studied. The main contents are given as follows:1. Classifying typical super-resolution algorithms, we introduce the methods how to lighten the burden from computational complexities and research for an optimal sparse construction. Simulations verify that algorithms based on subspace and sparse representation are all successful in achieving super-resolution. The former has a stronger immunity to the low SNR environment; Excluding the higher resolution, the latter gives up the coherency requirement between sources under the proper SNR.2. In order to get rid of the obstacles make subspace algorithms fail in super-multiple sources environment, a DOA estimation method based on a novel virtual array extension (VAE-FOCUSS) is proposed in the thesis. A special matrix is used to eliminate the redundancies of the covariance-like matrix firstly. Next, the virtual sensors are introduced by vectorizing and a mirror array. The total number of virtual sensors in the new idea is nearly the four times that of the physic array. The last step is to estimate DOAs through FOCUSS. The experiment results show that if enough snapshots are given, the VAE-FOCUSS algorithm will still work well without the need of noiseless conditions.3. The stronger the source correlation is, the weaker the orthogonality between the signal subspace and noise subspace is. Especially when coherent sources occur, the orthogonality disappears thoroughly. For the (Test of Orthogonality of Projected Subspaces, TOPS) algorithm, it’s a disaster. Hence, combining modified spatial smoothing technique, we offer a modified TOPS for the coherent wideband sources. To make full use of the array aperture and wideband information, a wideband DOA estimation algorithm with multiband alliance based on Orthogonal Match Pursuit (OMP) is suggested. Simulations show that qualities of the two proposed methods are both better than those of the traditional algorithms.