Study On DOA Estimation In The Presence Of Strong And Weak Signals

In the presence of strong and weak signals, the direction of arrivals (DOAs) of the weak signals cannot be estimate exactly because of the suppression and cover of the strong signals. For this problem, some research has been done. The main contributions and innovations are summarized as follows:1. Take a strong signal and a weak signal as an example, the separation of strong and weak signal after eigenvector transforming is deducted. And a conclusion that signals were distributed in different eigen beam by the descending order of the eigenvalues after eigenvector transforming is obtained.2. In the case of given source number, a new method based on eigenbeamforming normalization is proposed. This method is similar to MUSIC. The difference is that the seek direction vector should be pre-transformed by the normalized eigenvector matrix. The purpose is to make the power of signal to be identical which can slack the impact of the strong signals on the weak ones. Then the new signal space can be spanned by the unit vectors. Compared with the existing methods, this method has the advantages of slight computation burden and high estimation precision.3. In the case of unknown source number, a new method based on eigen beamforming for estimating source number and DOAs is proposed. Firstly, the eigenvectors of the spatial covariance matrix is sorted according to the descending order of the eigenvalues. Secondly, the spatial spectrum of each eigen beam is estimated in turn by the proposed equation. Finally, whether the signal source have or not is judged by the ration of the maximum value of the spatial spectrum to the mean value of the side-lobe peaks and the direction of arrival of the signal is estimated by the angle corresponding to the maximum value. The effectiveness and the superiority are demonstrated by the simulation results and the measured data.4. A new calibration method for a strong signal is proposed in the presence of array gain and phase mismatch. Firstly, the actual direction vector of the strong signal can be estimated by the eignvector according to the maximum eigenvalues. And the gain can been obtained. Then the strong signal can be estimated through the fitting of actual direction vector and gain multipled by seek vector. At last, the gain and phase can be estimated by the actual and theoretical vector of the strong signal.