Research On Oblique Projection Based Super-resolution Direction Finding Algorithm In Complex Signal Environment

The environment for propagating electromagnetic waves has been becoming complicated and diversified.The method of achieving direction of arrival(DOA)parameters of transmitting signals in complex environment has been a main research content of modern signal processing.In this subject,the problem of super-resolution direction estimation in complex signal electromagnetic environment is mainly concerned.And the super-resolution algorithms of estimating DOA with arbitrary passive receiving sensor arrays are discussed under unequal power signal environment,hybrid polarimetric signal environment and coherent signal environment using a technique which combines adaptive filtering and sparse signal reconstruction with spatial spectrum estimation.The main research aspects are summarized as following:Firstly,the mathematical models of scale and vector antenna arrays for receiving electromagnetic waves are discussed,and the oblique projection operator based spatial filtering algorithm,as well as the polarimetric signal filtering algorithm,in complex signal environment is researched.These results are the bases of the later chapters which aim at making a combination of adaptive filtering and super-resolution direction finding.To obtain the target polarimetric signal which impinging from interested angular direction with different contaminating noises,a generalized oblique projection operator based polarimetric signal filtering algorithm,which is based on multi-linear constrained interference suppression minimum variance criterion and can be used in complex signal environment,is proposed.The proposed algorithm adaptively suppresses receiving signals from undesired directions,and can be used to obtain signals from arbitrary expected directions.Also,it can be viewed as a further generalization and extension of the existing generalized oblique projection based spatial filtering algorithm in the polarization domain.Both theoretical analyses and simulation results indicate that the proposed algorithm constrains the array outputs to have the maximum output signal to interference and noise ratio and signal to interference ratio at the same time.Moreover,the proposed algorithm is effective for suppressing undesired signals and has strong robustness against certain derivations of a prior expected angular direction.Secondly,the influence of unequal power signal environment on super-resolution direction finding is analyzed.It is pointed out that the spatial spectrum peaks of weak signals could be sheltered by that of the adjacent strong ones when the conventional multiple signal classification(MUSIC)algorithm is used for DOA estimation.Herein,the generalized oblique projection spatial filtering approach is used to separate expected signal from the array measurements,and a spatial spectrum,which shows the power of each expected signal in the spatial domain,is constructed.Through adding weights to the spatial spectrum,the spectrum peaks of weak and strong signals are balanced.In this way,a synchronous processing of adapting filtering and spatial spectrum estimation is realized,the super-resolution DOA estimation of weak and strong signals is achieved,and the problem that weak signal being overwhelmed by strong signal is solved.Both theoretical analyses and simulation results show that the proposed algorithm does not depend on a priori location information of the incident signals,and it has robust performance for super-resolution DOA estimation of weak signals in unequal power signal environment.Thirdly,the influence of hybrid polarimetric signal environment on super-resolution direction finding is analyzed.It is pointed out that the conventional MUSIC has a poor performance or completely fails when it is used for DOA estimation of signals,which have time varying polarizations.By exploiting the spatial sparsity of incident signals,an oblique projection based polarization insensitive direction estimation algorithm is proposed.The proposed algorithm uses oblique projection based polarimetric signal filtering operators,which are insensitive to the polarization information of incident signals,to obtain estimates of receiving signals.The super-resolution DOA estimates are obtained via recognizing the spatial sparse structure of these separated signals.Both theoretical analyses and simulation results indicate that the proposed algorithm has excellent performance in estimating DOAs of arbitrary hybrid completely polarized and partially polarized signals.Moreover,it is independent of special polarimetric array configuration,and is applicable to hybrid polarimetric signal environment where the signals may have unequal power.Lastly,the influence of coherent signal environment on super-resolution direction finding is analyzed.The conventional MUSIC is shown to be ineffective for DOA estimation of coherent signals,no matter a scale or vector antenna array is used.Using oblique projection filtering and the relaxed optimization principle based iterative least squares,an alternating oblique projection based super-resolution DOA estimation algorithm is proposed to estimate DOA iteratively.In each iteration of the proposed algorithm,the DOA estimates,which are obtained in the previous iteration,are used as initial values in current iteration so that the oblique projection filtering operators can be constructed.Then these oblique projectors are alternatively used to separate incident signals into multiple single signals,and the DOAs of separated signals are estimated separately.The research results prove that the proposed algorithm is not restricted by the array geometry,and is able to estimate DOAs of noncoherent and coherent signals without exploiting decorrelation preprocessing.Moreover,its estimation accuracy can be asymptotically close to the Cram?er-Rao bound under certain conditions.To alleviate its high computational complexity,the proposed algorithm is combined with the alternating projection based maximum likelihood direction estimation method,and a new algorithm,which has lower complexity but keeps the same estimation accuracy,is developed.The simulation results verify the effectiveness of proposed algorithms.Additionally,both theoretical analyses and simulation results indicate that the proposed algorithms are also applicable for unequal power coherent signals and able to estimate DOAs of signals which have different degrees of polarization when there exist polarized coherent signals.The aforementioned research contents and achievements can be widely used in wireless communication,radar,electronic confrontation and many other areas which need to find radiation source locations.Also,they have certain practical significances in promoting the development and progression of super-resolution DOA estimation algorithms.