Research On Multi-parameter Estimation In Array Signal

Array signal processing systems have been used more and more widely in radar,communication and many other fields.The parameters of spatial source can be estimated by processing array's output.Compared with the single parameter estimation methods,multiparameter joint estimation methods are more suitable for practical application environment.In this thesis,the methods of joint angle and frequency estimation are researched on the problem of multi-parameter estimation.The research work of this thesis can be concluded as follows:Firstly,The basis of theorey of array signal processing is introduced.The model of array signal processing is reasearched under narrowband and broadband conditions respectively.The 2nd order statistical characteristics of array model is analyzed for the base of parameter estimation.Secondly,The methods of signal frequency estimation are studied.The narrowband signal frequency estimation algorithms based on undersampling are studied in wideband array.It avoid the problem of massive data storage and the underlying difficulty in signal processing.The methods of frequency estimation for broadband signal are studied.The algorithm based on STFT and least square fitting and based on Fr FT and peak searching respectively applied to estimate the frequency parameters of the broadband LFM signal.Thirdly,The methods of direction of arrival(DOA)estimation are studied.The Conventional Beam Forming(CBF)algorithm is only a simple expansion of Fourier spectrum estimation to the airspace,and its angular resolution is restricted by the Rayleigh limit.The classical subspace algorithm: MUSIC and ESPRIT is studied respectively.To avoid searching the spectral peak in MUSIC,the Root-MUSIC algorithm is introduced as the optimization of MUSIC.Under the broadband condition,the ISM algorithm for incoherent signals and the improved ISM algorithm,CSM algorithm for coherent signal are studied respectively,and the related simulation and verification have been completed.Finally,The joint DOA and frequency estimation methods are studied.Under the narrowband condition,in order to avoid the great computation complexity in 2D MUSIC, the ESPRIT algorithm based on the smoothing technology is studied.Combined with the smoothing technology,The algorithm constructs a data model which conforms to the ESPRIT by the array element output,which don't use delay channel.The DOA and frequency of multiple narrowband signals can be estimated by the rotation matrix,which can be obtained by the method of joint diagonalization.Under the broadband condition,The method of joint DOA and frequency estimation for wideband LFM signal is studied.The algorithm estimates frequency parameters by the WVD of reference array's output and Hough transformation.Then the array data matrix in time-frequency domain can be constructed by processing the array's output with WVD and extracting its time frequency ridge.The DOA can be estimated by applying space spectrum estimation method to the correlation matrix that in time-frequency domain.Therefore,the joint DOA and frequency estimation of LFM signal is realized.