Research On Ground Moving Target Detection And Spurious Target Recognition For Multichannel Airborne Radar

As airborne synthetic aperture radar(SAR)has superior azimuth resolution,range resolution and stong maneuverability,it is one of the popular radar systems in recent years.The long-time synthetic aperture in SAR systems facilitates the increasement of signal-to-noise ratio of groud moving targets such that weak and slowly target can be detected.Therefore,the combination of SAR and ground moving target can not only obtain high-resolution scene image,but also detect weak and slow moving target,estimate the parameters of moving target,and relocate the target in the interested area.In the practical application of multichannel SAR-GMTI,the problem of moving target detection is studied by using advantage of multichannel degree of freedom(DOF).In this dissertation,we focuse on the clutter suppression in heterogeneous environments,ground moving target indication in the area of interest,the robust parameters estimation of moving targets,the problem of the range ambiguity under long-distance detection,the influence of spurious targets on the recognition probability of real targets and so forth.The main contents of this dissertation are summarized as follows.1.Robust principal component analysis(RPCA)is an important clutter suppression method for synthetic aperture radar-ground moving target indication(SAR-GMTI).However,the phase difference among channels caused by the velocity of moving target is not sensitive for the correlation coefficient.It results in that the conditions of applying RPCA method cannot be satisfied.Moreover,the azimuth velocity of the moving target cannot be estimated since the phase information is destroyed by using the soft-thresholding operator.To solve these problems,a clutter suppression method based on WVD-RPCA is proposed in chapter two.By analysis,the phase difference among channels is transformed to the shifting in time-frequency domain by using WVD transformation.Therefore,the correlation coefficient between channels is decreased,which means that the sparse matrix of moving targets can be extracted with higher probability.But the velocity of moving targets cannot be estimated using the proposed method.Then,to overcome the shortcoming of the above proposed method,a novel phase envelope transform-robust principal component analysis(NPET-RPCA)method is proposed.First,the modified correlation kernel function(MCKF)of the echo data is constructed to decoupling between the slow time and lag-time.Second,the phase difference among channels caused by target motion is transformed into envelope delay by generalized scaled Fourier transform(GSCFT)and fractional Fourier transform.Thus,the correlation coefficient between channels for moving target is decreased.At last,the moving targets are separated by the RPCA in the Doppler chirp rate-transform domain.Additionally,the echo data is focused in the Doppler chirp rate-transform domain such that the azimuth velocity could be estimated by the positions of the focused points.Simulation results illustrate the effectiveness of the proposed method.2.In general,it is hard to acquire sufficient secondary data in heterogeneous environments.To solve the problem in clutter suppression,a method based on nonlocal self-similarity-robust principal component analysis(NSS-RPCA)is proposed for airborne SAR systems.First,discrete clutter is separated from the echo data by RPCA and division of similar blocks after range pulse compression.Second,similar blocks of the residual-clutter background are extracted to overcome the training sample limitation using the NSS method in the two-dimensional time domain.Third,subcovariance matrices are structured by the similar blocks,and the subcovariance matrix is stacked into a tensor.Then,the subclutter covariance matrix can be obtained from the stacked subcovariance matrix tensor by RPCA,where the residual clutter tensor is of low rank and the target tensor is sparse.Finally,the residual clutter can be suppressed by the subclutter covariance matrix.In this manner,the source of independent identically distributed(IID)samples will be increased significantly without aperture loss by the proposed method.Simulation and analysis based on the experimental data illustrate the effectiveness of the proposed method.3.In the traditional moving targets detection based on RPCA for multichannel SAR systems,the nonsparsity of the moving target echoes limits the performance of the RPCA in SAR-GMTI,and the velocity of the moving target cannot be estimated since the phase information is destroyed by the soft-thresholding operator in the RPCA process.To solve these problems,a novel moving target detection method based on RPCA(NRPCA)for multichannel SAR systems is proposed in chapter four.An atomic norm-based optimization program is first constructed to transform the data sparsity requirement into a moving target sparsity requirement.Although this optimization program is NP-hard,it can be transformed to semidefinite programming by relaxation.Furthermore,accurate velocity estimation is performed using dual function theory and the alternating direction method of multipliers(ADMM)algorithm,while selection of the sparsity order k is avoided.Simulations and analyses based on experimental data illustrate the effectiveness of the proposed method.4.In multi-channel SAR moving target detection systems,the image coregistration and channel phase errors deteriorate the performance of the moving target radial velocity estimation.To solve this problem,a robust radial velocity estimation method is proposed using a frequency diverse array-synthetic aperture radar(FDA-SAR)model in chapter five.By introducing the step frequency,the Doppler frequency varies along different channels and the valuable phase difference becomes a linear function of Doppler frequency.The radial velocity of moving targets is embedded in the first-order term which does not include error terms.Therefore,the accurate velocity of moving targets is estimated by the first-order coefficient which is solved by least-squares fitting method.Owing to the analysis and derivation,the proposed method is robust on the condition of image coregistration error.Simulations and data analysis illustrate the effectiveness of the proposed method.5.Under long-distance detection,the range ambiguity leads to the echo of the near ground enter the receiver through the sidelobe,while the reflected echo of the area of interesting(AOI)enter receiver by the mainlobe,resulting in poor imaging quality for SAR systems.To solve the problem of the range ambiguity,a mitigating range ambiguities based on Doppler division multiple access(DDMA)method is proposed in chapter six.First,after the pulse compression and phase compensation,the echo data between the different transmitters are processed by interferometric method.Then,the coupling phase term between the ambiguous range and the offset Doppler caused by DDMA waveform is generated.Utilizing the coupling phase term,the echo of the different ambiguous areas can be distinguished in the slow time domain.Thereafter,by designing passband filter in the slow time domain,the power of the ambiguity region is suppressed in the unambiguity region.Finally,the result of the supeior imaging quality is obtained by the proposed method.Simulation results exhibit the verified effectiveness of the proposed method.6.In target indication,if there are spurious targets in the detected targets,the false alarm rate of radar system will increase.In order to solve the two problems that the error is caused by omitting the high-order terms of Taylor expansion and the recognition probability of spurious target is not satisfactory by using unconstrained statistical hypothesis test,a constrained statistical hypothesis test based on MIMO-SAR method is proposed to recognize spurious and real targets.First,utilizing the characteristic of bistatic MIMO radar,the improved equations are constructed for the bistatic range history,the transmit cone angle and the receive cone angle.Second,a nonlinear transformation is introduced to overcome the error caused by omitting the high-order terms of Taylor expansion in traditional method.Third,the real and spurious targets are recognized under the constrained F hypothesis test instead of^?2 test.Compared with the traditional method,the proposed method exhibits superior recognition probability.At last,simulation results are given to demonstrate the effectiveness of the proposed method.