Research On Autofocus Algorithm Forbistatic Forward-looking SAR

In the field of autonomous landing and autonomous navigation,there is a strong demand for radar forward looking imaging capability.By setting the transmitter and receiver on two separate platforms,bistatic SAR can form quasi orthogonal of equal distance line and equal doppler line in front of the receiving station.Hence,it has the potential of forward looking imaging,and it has become a hot spot of research.However,compared with monostatic SAR,bistatic SAR has greater freedom of movement and more error sources,and the compensation of motion error is more important.Autofocus refers to the method of extracting motion errors from the echo data and compensating them,which is the key step of the bistatic forward looking SAR(BFSAR)to get high resolution images.In this paper,we focus on the following aspects: echo modeling,error effect,parameter equalization,error extraction and so on,which are related to the autofocus of BFSAR.1.The geometry model and the echo signal model of BFSAR are established.The expression of range history of the point target is derived.Under the ideal condition,the slant-range law and Doppler signal characteristics of BFSAR are analyzed.The influence of the motion error of BFSAR is simulated and analyzed.This work lays the theory for the follow-up autofocus algorithm.2.In view of the azimuth variance of Doppler parameter in BFSAR,which is introduced by linear RCM correction(LRCMC),and the law of azimuth variance is analyzed.Then the azimuth signal is modeled as a multicomponent polynomial signal(mc-ps).Parameter estimation of mc-ps based on the modified WVD is studied.And the work of this part provides the precondition for the autofocus scheme.3.Nonlinear chirp scaling(NLCS)algorithm under non-ideal condition is studied.On the basis of estimating the parameters of the azimuth mc-ps,the analytical form relationship between Doppler parameters and azimuth coordinates will be established by least square fitting.Then,the perturbation function is constructed and the azimuth variance under the non-ideal condition is equalize by the NLCS theory.4.On the basis of equalized azimuth variant Doppler parameters,a phase error extraction method based on image contrast optimization is studied.Firstly,coordinate descent method is used to convert the search process of phase error in high dimensional space into one dimension search for phase errors of each azimuth unit,and then the one dimension search problem of phase error is transformed into four order polynomial root problem by multidimensional projection method.Compared with the existing methods,the proposed method has no order restriction and greatly improves the computation efficiency.The methods involved above have all passed the simulation test.