Reasearch On Subband Fusion Techniques For Ultra Wideband

Subband fusion is a novel technique for radar signal processing,which uses the subband data from multiple radars operating in different frequency bands to fuse anultrawideband signal.It can effectively improve the resolution of radar without enhancing the hardware of original radar system.In this paper,a research on subband alignmenttechniques for motion target and subband fusion problems in complex noise environment is presented,specific tasks are as follows:1.Most aerial targets are movable in practical application.Firstly,an echoed signal model of motion target based on geometrical theory of diffraction(GTD)is established.After the influence of velocity on subband alignment algorithm is analyzed,amethod based on minimum entropy principle is used to compensate velocity.Then,the definition and content of incoherent factors are given.Next,the subband alignment method based on time domain correlation and the subband alignment method based on the pole estimation value are researched separately.It is shown in the simulation,when Signal-toNoise Ratio(SNR)is high,both of them can estimate the phase factors well and make the subbands coherent.2.The estimation precision of the former method is limited by the errors of bandwidth extrapolation(BWE),while model ordering of the latter method is susceptible to noise.Therefore,a subband alignment method based on sparse reconstruction is proposed,which takes the sparseness of echo signals in the frequency domain after pulse compression into account and Sparse Bayesian Learning(SBL)method is used to reconstruct the one-dimensional range profile.Then,the linear factor of phase difference is estimated using the positional relationship among one-dimensional range profiles of subbands and the fixed phase is obtained by solving the nonlinear least squares problem.Finally,the amplitude difference of subbands can be evaluated by the amplitude ratio of the higher band and lower band.In the simulation experiment,It is shown that the proposed method has higher recovery accuracy of coherence,compared with two algorithms,when is SNR low.3.Real noise environment may not be Gaussian,the subband fusion problem in complex noise environment is proposed.Firstly,according to the probability density function(pdf)of different noise environments,corresponding penalty functions can be established.It is shown that the Maximum A Posteriori probability estimate can be obtained by iteratively solving a support vector regression problem,and the full frequency band can be reconstructed by the estimated parameters.It is verified that the proposed algorithm outperforms the existing algorithms in complex noise environment based on computer simulation results,FEKO simulation results and the static range data measured in the microwave darkroom.