Off-Grid Targets Localization Algorithm For Frequency Diverse Array Radar Based On Sparse Bayesian Learning

Targets localization technology based on DOA(direction of arrival)estimation is one of the important contents of array signal processing research,which has been widely used in radar field.The traditional phased array radar beam pointing is fixed at an Angle over all ranges.So it is impossible to estimate the target's range information directly from its beamforming output due to the inherent range ambiguity.The frequency diverse array(FDA)radar is different from the traditional phased array radar in that its beampattern is angle-range dependent,so it is widely used in the angle range two-dimensional localization of radar targets.In this thesis,FDA radar is applied to targets localization to solve the problem that phased array radar has range ambiguity and can't achieve target angle range location effectively.The main work of this thesis is summarized as follows:1.To deal with the problem that the traditional subspace based DOA estimation methods are constrained by small snapshots,low signal-to-noise ratio(SNR)and correlated signals.Introducing the SBL algorithm into DOA estimation,considering that SBL mainly uses real signals for research,while the signals are complex in DOA estimation,and the results obtained in real model cannot be directly applied to complex model.Therefore,this paper studies the extension of SBL from the real model to the complex model,and unifies the expression of the SBL theory under the two models by the value of the parameters.This form enables to generalize and improve the sparse Bayesian methods proposed nowadays.2.Combined with the complex model of DOA estimation,a complex multisnapshot SBL for off-grid DOA estimation method is proposed.Firstly,using the spatial sparsity of signals,the DOA estimation problem is transformed into sparse signal reconstruction(SSR)problem.secondly,considering the problem of insufficient estimation accuracy caused by spatial grid sampling mismatch in sparse model,an offgrid DOA estimation model is introduced,and grid offset is introduced.Finally,SBL algorithm is used for sparse reconstruction to realize off-grid DOA estimation.The simulation results show that this method can effectively process complex signals by SBL,and complete the off-grid DOA estimation of complex signals with ultra-high accuracy.3.In view of the existing FDA radar target location methods,most of them are traditional beam scanning and subspace methods,which are computationally expensive and difficult to implement.In this thesis,SBL combined with FDA radar characteristics is used for targets angle range two-dimensional localization,and a double-pulse FDA radar off-grid targets localization method based on SBL is proposed.Firstly,log frequency offset is used at the transmitter to remove the range periodicity of the beam pattern of the full-band frequency receiving mechanism of FDA radar for range unambiguous localization.Secondly,the difficulty of applying SBL directly to angledistance two-dimensional estimation is analyzed,and a double-pulse method is considered.Transmit zero frequency offset and non-zero frequency signals through the FDA radar to convert the angle-range two-dimensional estimation into two onedimensional estimates of the angle and range respectively,reducing the amount of calculation.The simulation results show that the SBL combined with the characteristics of FDA radar can achieve the requirements of targets localization with utra-high accuracy,which shows its superiority over the traditional algorithm.