Direction-of-arrival and frequency estimation in array signal processing based on blind source separation
Wen-Shu Xiao,Xing-Gan Zhang
DOI: https://doi.org/10.3321/j.issn:0469-5097.2009.04.005
2009-01-01
Abstract:In recent years,blind signal separation(BSS) has already become a hot field of signal processing and neural network,but it has been mainly applied to speech signal processing,and the application of BSS to radar signal processing is rare.Estimating the direction-of-arrival and frequency is the basic issue of array signal processing.Recently,BSS has already become a hot field of signal processing and neural network.Over the past 20 years,different algorithms withdifferent performance have been developed,such as maxi mumlikelihood,maxi mumentropy and eigen decompositionbased on autocorrelation matrix.Especially,the MUSIC and ESPRIT are devised to deal with direction-of-arrivalesti mation(DOA).In this year,some researchers apply the BSS to the DOA,which is a new method to esti mate the DOA andfrequency,but the methods introduced in these literatures all have li mitations,for example :the convergence of thealgorithmis shown to be slow,and it is not robust,etc.Based on the methods which are proposed by some literatures,firstly,in this paper,the model of the signalsreceived by radar array has been studied,and the attributes of the received signals of radar array(the combination ofradar clutter and the echo of the object) have also been analyzed.Then a new BSS algorithmfor complex signals hasbeen proposed.Finally by applying the new BSS algorithm,a newesti mation algorithm,which can both esti mate theDOAand frequency,has been proposed in this paper.The new BSS algorithm is based on the fast fixed-point independent component(fast-ICA),and does notdepend on any user-defined parameters.So the new BSS algorithmis fast to converge to the most accurate solutionallowed by the data and the algorithm.In this paper the convergence of the algorithmis rigorously proven,and theconvergence of the algorithmis shown to be cubic.The si mulation results showthat :1) The new DOAesti mation algorithmis efficient and robust.2) The algorithmcan esti mate the direction-of-arrival and frequency fromthe source signals with noise.3) The algorithmcan also separate the radar clutter and the echo of the object fromthe received signals of radararray.