Sequential Direction Of Arrival Estimation Algorithms For Coherent Signals

Direction-of-arrival(DOA)estimation is an important research branch of the array signal processing,and it has been widely used in radar,communication,sonar,conference system and so on,which makes it develop quickly for the past fifty years.In real environment,there are a lot of coherent signals,because of the multipath effect caused by the reflect of mountains and buildings,and military man-made electromagnetic interference.All coherent signals are estimated simultaneously by conventional decorrelation algorithms,which will largely lose the array aperture,and will restrict the number of estimated coherent signals and the DOA estimation accuracy to a great extent.Moreover,there are all kinds of error from array antenna in real application,which leads to serious decline on the DOA estimation performance.Therefore,researching DOA estimation algorithms in the presence of array error not only has very important theoretical significance but also practical value.In this paper,we focus on the grouping DOA estimation for coherent signals on the uniform linear array,and several effective algorithms are proposed.The main works are summarized as follows:1.An approach based on forward spatial smoothing is presented to estimate the DOAs of coherent signals sequentially group by group,which runs in turn according to the number of coherent signals in each group.We first estimate the DOAs of one group which has the minimum coherent signals,and the fading coefficients of this group are estimated,and the information of this group is eliminated from the array covariance matrix with oblique projection technique.Then we estimate the next group similarly until all groups are estimated.Furthermore,the improved algorithm based on the forward/backward spatial smoothing is proposed to further improve the estimation accuracy of the DOAs in each group separately.The components of each group can be extracted separately from the array covariance matrix by oblique projection technique,and the estimation accuracy of each group is further promoted using more subarrays.Simulation results are presented to demonstrate the proposed method has better performance in low signal-to-noise ratio,small number of snapshots and more signals.2.We propose an algorithm which is to estimate the DOAs of coherent signals sequentially group by group in the presence of mutual coupling.Taking the advantage of the feature of uniform linear array,and the relationship between mutual coupling and distance,the elements at the both end of array are treated as auxiliary ones.The elements in the middle array have the same mutual coupling environment,and the mutual coupling can be equivalent to the fading coefficient,which can ignore the influence of mutual coupling.Therefore,the algorithm based on forward spatial smoothing that is to estimate the DOAs of coherent signals sequentially group by group can be used on the middle array,which can resolve the DOAs of coherent signals sequentially group by group in the presence of mutual coupling.3.An algorithm is presented to estimate the DOAs of independent and coherent signals in the presence of mutual coupling.Making use of the feature that independent signals are orthogonal to noise subspace without spatial smoothing,the mutual coupling can be equivalent to signal energy in the middle array.So,The DOAs of independent signals can be estimated firstly,and then the fading coefficients can be resolved using the estimated DOAs of independent signals.Afterwards,the information of independent signals is eliminated from the array covariance matrix and the mutual coupling is removed,and then coherent signals can be estimated group by group.