Research On Adaptive Beamforming And DOA Estimation Techniques Of Virtual Antenna Array Based On Fourth Order Statistics

Adaptive beamforming and direction of arrival(DOA)estimation are both array signal processing techniques.They have many important applications in the field of military and civil wireless communication.Most of the adaptive beamforming methods of array antenna need accurate DOA information of the desired signal,so it is closely related to the DOA estimation technique.The results of DOA estimation can be used as the priori information of adaptive beamforming to improve its output performance.However,the number of actual array elements and the size of the array antenna aperture will limit the performance of the conventional adaptive beamforming and DOA estimation methods.When the number of sources received by the array antenna exceeds the degree of freedom(DOF)of the actual array antenna,the performance of the conventional adaptive beamforming and the DOA estimation method will greatly degrade.The virtual antenna array(VAA)technology based on fourth order statistics can increase the DOF of the array antenna and extend the aperture of the array effectively by constructing virtual array element,and the performance of beamforming and DOA estimation can be improved.The DOA estimation of the VAA based on fourth order statistics is studied,which improves the resolution performance of DOA estimation,and provides more accurate DOA estimation results for adaptive beamforming as the priori information.Then,we focus on adaptive beamforming of the VAA based on fourth order statistics.The aim is to improve the output performance of the array antenna beamforming.The main contents of this paper are as follows:Aiming at the problem that the performance of array antenna adaptive beamforming can be improved if more accurate DOA of the desired signal can be obtained,two DOA estimation approaches of the VAA are proposed to improve the resolution of DOA estimation,by which more accurate DOA estimation results can be obtained.Firstly,the reconstructed fourth order statistics based multiple signal classification(RFO-MUSIC)approach is proposed by reconstructing the fourth order covariance matrix.The resolution of DOA estimation is improved without increasing the computational load compared with the conventional FO-MUSIC approach.Then,the twice virtual extension multiple classification(TVEM)approach is proposed.The conjugate virtual extension technology is combined in order to realize the twice virtual extension.Compared with the existing DOA estimation methods of the VAA,higher resolution of DOA estimation can be obtained by the TVEM approach.Aiming at the problem that the computational load of the VAA beamforming is too large to be used in the fast beamforming situations like the conventional adaptive beamforming,the semi virtual antenna array(SVAA)beamforming approach is proposed.The covariance matrix is expanded by the Kronecker product of the eye matrix and the sampling covariance matrix.Virutual array elements are formed and more interference than the DOF of the array can be inhibited by constructing the linear constraints with zero response.Besides,the fast calculation method of the weight vector is derived according to the mathematical properties of the eye matrix and the Kronecker product.The computational load of the SVAA beamforming approach is close to that of the conventional adaptive beamforming approaches.Aiming at the problem that the performance of null broadening beamforming is limited when the number of array elements is small in the case of fast moving interference,two null broadening beamforming approaches of VAA are proposed.Virtual array elements are formed and the performance of null broadening beamforming is improved,by which fast moving interference can be better handled.Firstly,the fourth order statistics based covariance matrix taper(FO-CMT)approach is proposed by reconstructing the fourth order interference plus noise covariance matrix.The covariance matrix taper(CMT)approach is also combined.The performance of null broadening beamforming is improved.Then,the SVAA beamforming approach is combined with the linear constraint sector suppressed(LCSS)null broadening beamforming approach,and the covariance matrix expansion based LCSS(CME-LCSS)null broadening beamforming approach is proposed.Virtual array elements are formed,and more linear constraints with zero response are constructed.Compared with the LCSS approach,which is the most advanced null broadening beamforming approaches,better performance can be obtained by the CME-LCSS approach.Therefore,fast moving interference can be better suppressed and the output performance of beamforming is improved.Aiming at the problem that the performance of the existing null broadening beamforming approaches will seriously degrade when there are array calibration errors,a null broadening beamforming approach of VAA against array calibration errors is proposed.The uncertainly quadratic optimization and diagonal loading techniques are combined by the VAACE approach.The sampling covariance matrix and the steering vector of the array are expanded.Then,linear constraints with zero response are constructed according to the corrected steering vectors.Compared with the existing methods,the null broadening beamforming performance of the VAACE method is greatly improved.The robustness of the VAACE approach to array calibration errors is also improved.Therefore,jammer motion can be better suppressed and the output performance of beamforming is improved.