Email alert
2016 Vol. 38, No. 12
Display Method:
2016, 38(12): 2988-2995.
doi: 10.11999/JEIT161034
Abstract:
To improve the performance of radar High-Resolution Range Profile (HRRP) target recognition, a new attention-based model is proposed based on time domain feature. This architecture encodes the time domain feature which can reveal the correlation inside the target with Recurrent Neural Network (RNN). Then, this model gives a weight to each part and sums the hidden feature with each weight for the final recognition. Experiments based on measured data show that the attention-based model is effective for radar HRRP recognition. Furthermore, the proposed method can still find the support areas even with the removed test data.
To improve the performance of radar High-Resolution Range Profile (HRRP) target recognition, a new attention-based model is proposed based on time domain feature. This architecture encodes the time domain feature which can reveal the correlation inside the target with Recurrent Neural Network (RNN). Then, this model gives a weight to each part and sums the hidden feature with each weight for the final recognition. Experiments based on measured data show that the attention-based model is effective for radar HRRP recognition. Furthermore, the proposed method can still find the support areas even with the removed test data.
2016, 38(12): 2996-3003.
doi: 10.11999/JEIT161068
Abstract:
Feature extraction is a key step and difficult point in SAR image target recognition. This paper presents a novel method based on Poisson Gamma Belief Network (PGBN) for SAR image target recognition. As a deep Bayesian generative network, the PGBN model obtains a more structured multi-layer feature representation from the complex SAR image data using the high nonlinearity of the Gamma distribution, and the multi-layer feature representation effectively improves SAR image target recognition performance. In order to obtain a higher recognition rate and efficiency of training, this paper further proposes a method for classifying PGBN model based on the Naive Bayes rule. The experimental results about MSTAR dataset show that the feature extracted by this new method has better structure information, and it has better performance for SAR image target recognition.
Feature extraction is a key step and difficult point in SAR image target recognition. This paper presents a novel method based on Poisson Gamma Belief Network (PGBN) for SAR image target recognition. As a deep Bayesian generative network, the PGBN model obtains a more structured multi-layer feature representation from the complex SAR image data using the high nonlinearity of the Gamma distribution, and the multi-layer feature representation effectively improves SAR image target recognition performance. In order to obtain a higher recognition rate and efficiency of training, this paper further proposes a method for classifying PGBN model based on the Naive Bayes rule. The experimental results about MSTAR dataset show that the feature extracted by this new method has better structure information, and it has better performance for SAR image target recognition.
2016, 38(12): 3004-3010.
doi: 10.11999/JEIT160905
Abstract:
For the problem of detection performance loss of adaptive detectors on the condition that the secondary data are limited, the adaptive detection method of range-spread target based on the prior knowledge of clutter is proposed. The texture and the covariance matrix of speckle of clutter are respectively modeled as the random variable which follows the inverse Gamma distribution and the random matrix which follows the inverse complex Wishart distribution. Based on the prior knowledge, the Maximum A Posteriori (MAP) estimation of texture component is obtained and the adaptive detector of range spread target which does not need the secondary data is designed via utilizing the generalized likelihood ratio test. Finally, the detection performances of the proposed detector are evaluated and the experimental results illustrate that the proposed detector is robust in parameters mismatched situation and outperforms the conventional generalized likelihood ratio test detector for range-spread target in limited secondary data scenarios.
For the problem of detection performance loss of adaptive detectors on the condition that the secondary data are limited, the adaptive detection method of range-spread target based on the prior knowledge of clutter is proposed. The texture and the covariance matrix of speckle of clutter are respectively modeled as the random variable which follows the inverse Gamma distribution and the random matrix which follows the inverse complex Wishart distribution. Based on the prior knowledge, the Maximum A Posteriori (MAP) estimation of texture component is obtained and the adaptive detector of range spread target which does not need the secondary data is designed via utilizing the generalized likelihood ratio test. Finally, the detection performances of the proposed detector are evaluated and the experimental results illustrate that the proposed detector is robust in parameters mismatched situation and outperforms the conventional generalized likelihood ratio test detector for range-spread target in limited secondary data scenarios.
2016, 38(12): 3011-3017.
doi: 10.11999/JEIT161072
Abstract:
For the detection issue in the presence of subspace signal mismatch, a parametrically tunable detector is proposed, which processes the Constant False Alarm Rate (CFAR) properties. By changing the tunable parameter, the proposed detector can flexibly detect the mismatched subspace signal. Moreover, in the case of no signal mismatch, the proposed tunable detector can even achieve better detection performance than the existing detectors. The expressions for the probabilities of detection and false alarm are proposed and verified with Monte Carlo simulations.
For the detection issue in the presence of subspace signal mismatch, a parametrically tunable detector is proposed, which processes the Constant False Alarm Rate (CFAR) properties. By changing the tunable parameter, the proposed detector can flexibly detect the mismatched subspace signal. Moreover, in the case of no signal mismatch, the proposed tunable detector can even achieve better detection performance than the existing detectors. The expressions for the probabilities of detection and false alarm are proposed and verified with Monte Carlo simulations.
2016, 38(12): 3018-3025.
doi: 10.11999/JEIT161032
Abstract:
This paper studies the issue of SAR target detection with CNN when the training samples are insufficient. The existing complete dataset is employed to assist accomplishing target detection task, where the training samples are not enough and the scene is complicated. Firstly, the existing complete dataset with image-level annotations is used to pre-train a CNN classification model, which is utilized to initialize the region proposal network and detection network. Then, the training dataset is enlarged with the existing complete dataset. Finally, the region proposal model and detection model are obtained through the pragmatic 4-step training algorithm with the augmented training dataset. The experimental results on the measured data demonstrate that the proposed method can improve the detection performance compared with the traditional detection methods.
This paper studies the issue of SAR target detection with CNN when the training samples are insufficient. The existing complete dataset is employed to assist accomplishing target detection task, where the training samples are not enough and the scene is complicated. Firstly, the existing complete dataset with image-level annotations is used to pre-train a CNN classification model, which is utilized to initialize the region proposal network and detection network. Then, the training dataset is enlarged with the existing complete dataset. Finally, the region proposal model and detection model are obtained through the pragmatic 4-step training algorithm with the augmented training dataset. The experimental results on the measured data demonstrate that the proposed method can improve the detection performance compared with the traditional detection methods.
2016, 38(12): 3026-3033.
doi: 10.11999/JEIT161038
Abstract:
By suppressing the Doppler ambiguity, the along-track multi-channel Synthetic Aperture Radar (SAR) system can simultaneously achieve High-Resolution and Wide-Swath (HRWS) imaging. However, the presence of unavoidable amplitude and phase bias tends to the absence of ambiguous signals in the SAR images. To address this issue, a novel phase bias estimation algorithm based on Doppler spectrum optimization is proposed. By exploiting the fact that phase bias can cause Doppler spectrum broadened, the phase bias can be successfully estimated by optimizing the Doppler spectrum. The Doppler centroid estimation can be avoided before phase biases estimation, which reduces the estimation accuracy caused by the inaccurate Doppler centroid. The proposed algorithm can achieve better performance when Signal to Noise Ratio (SNR) is low. The effectiveness of the algorithm is validated by experimental results carried out on simulated data and SAR data collected by an air- borne multi-channel system.
By suppressing the Doppler ambiguity, the along-track multi-channel Synthetic Aperture Radar (SAR) system can simultaneously achieve High-Resolution and Wide-Swath (HRWS) imaging. However, the presence of unavoidable amplitude and phase bias tends to the absence of ambiguous signals in the SAR images. To address this issue, a novel phase bias estimation algorithm based on Doppler spectrum optimization is proposed. By exploiting the fact that phase bias can cause Doppler spectrum broadened, the phase bias can be successfully estimated by optimizing the Doppler spectrum. The Doppler centroid estimation can be avoided before phase biases estimation, which reduces the estimation accuracy caused by the inaccurate Doppler centroid. The proposed algorithm can achieve better performance when Signal to Noise Ratio (SNR) is low. The effectiveness of the algorithm is validated by experimental results carried out on simulated data and SAR data collected by an air- borne multi-channel system.
2016, 38(12): 3034-3041.
doi: 10.11999/JEIT160058
Abstract:
To solve the contradiction between estimation performance and computation burden, when detecting high speed target in wideband noise radar, a novel method based on Frequency Domain Supper-Resolution (FDSR) is proposed for parameter estimation. Firstly, a group of component noise frequency modulation signals with different phase information are obtained by setting different fixed delay, and the Doppler phase difference group is constructed through matched filtering without scaling transform. Then according to the similarity of Doppler phase difference group and array signal processing, the velocity is acquired through modern spectrum estimation algorithm, and the range is calculated by scaled matched filtering after Doppler compensation. In this method, the velocity is obtained without considering Doppler dispersion and the time-domain reconstruction is used only once in the whole algorithm. The computation burden and estimation performance both superior to wideband cross ambiguity function algorithm based on curve fitting, when the fixed delay and group number of Doppler phase difference group are appropriately selected. The effectiveness of this algorithm is validated by simulation results.
To solve the contradiction between estimation performance and computation burden, when detecting high speed target in wideband noise radar, a novel method based on Frequency Domain Supper-Resolution (FDSR) is proposed for parameter estimation. Firstly, a group of component noise frequency modulation signals with different phase information are obtained by setting different fixed delay, and the Doppler phase difference group is constructed through matched filtering without scaling transform. Then according to the similarity of Doppler phase difference group and array signal processing, the velocity is acquired through modern spectrum estimation algorithm, and the range is calculated by scaled matched filtering after Doppler compensation. In this method, the velocity is obtained without considering Doppler dispersion and the time-domain reconstruction is used only once in the whole algorithm. The computation burden and estimation performance both superior to wideband cross ambiguity function algorithm based on curve fitting, when the fixed delay and group number of Doppler phase difference group are appropriately selected. The effectiveness of this algorithm is validated by simulation results.
2016, 38(12): 3042-3048.
doi: 10.11999/JEIT160859
Abstract:
This paper proposes a clutter suppression and parameter estimation method based on Relax algorithm. The method is fit for the multi-channel Wide Area Surveillance (WAS) GMTI system. After analyzing the components of radar echoes in WAS-GMTI mode, an iterative clutter suppression method based on Relax algorithm is designed. Compared with the reduced-dimension Space Time Adaptive Processing (STAP) algorithm, the proposed method achieves better clutter suppression results in the nonhomogeneous environment since it has no need to estimate the clutter plus noise covariance matrix. Besides, based on the clutter suppression results, precision parameters of moving targets can be obtained if the iterations are continued in the range-Doppler cells with the detected moving targets. The proposed method is validated by digital simulations.
This paper proposes a clutter suppression and parameter estimation method based on Relax algorithm. The method is fit for the multi-channel Wide Area Surveillance (WAS) GMTI system. After analyzing the components of radar echoes in WAS-GMTI mode, an iterative clutter suppression method based on Relax algorithm is designed. Compared with the reduced-dimension Space Time Adaptive Processing (STAP) algorithm, the proposed method achieves better clutter suppression results in the nonhomogeneous environment since it has no need to estimate the clutter plus noise covariance matrix. Besides, based on the clutter suppression results, precision parameters of moving targets can be obtained if the iterations are continued in the range-Doppler cells with the detected moving targets. The proposed method is validated by digital simulations.
2016, 38(12): 3049-3055.
doi: 10.11999/JEIT160922
Abstract:
To solve the problem of motion parameter estimation of ground moving target, a parameter estimation method of moving targets with sparse sampling data of single SAR sensor is proposed. First, based on the 2 dimensional velocity of moving targets, an equivalent parametric model is constructed to transform the moving target echo into squint SAR echo. Then, with different parameters the modified iterative thresholding algorithm is applied to achieving imagery of moving target. Finally, the motion parameters of targets are obtained by minimizing the image entropy. It is shown that, using the proposed method, the required echo sampling can be reduced, the Doppler ambiguity problem can be avoided and accurate velocity estimation can be obtained even in low signal-to-clutter ration scenarios. Simulation results verify the effectiveness of the proposed method.
To solve the problem of motion parameter estimation of ground moving target, a parameter estimation method of moving targets with sparse sampling data of single SAR sensor is proposed. First, based on the 2 dimensional velocity of moving targets, an equivalent parametric model is constructed to transform the moving target echo into squint SAR echo. Then, with different parameters the modified iterative thresholding algorithm is applied to achieving imagery of moving target. Finally, the motion parameters of targets are obtained by minimizing the image entropy. It is shown that, using the proposed method, the required echo sampling can be reduced, the Doppler ambiguity problem can be avoided and accurate velocity estimation can be obtained even in low signal-to-clutter ration scenarios. Simulation results verify the effectiveness of the proposed method.
2016, 38(12): 3056-3062.
doi: 10.11999/JEIT161036
Abstract:
Micro-Doppler (m-D) feature extraction is significant for group target discrimination, while the methods for single target are invalid. An m-D feature extraction method of group targets is proposed based on signal orthogonal decomposition. First, the Sinusoidal Frequency-Modulated (SFM) form of m-D signals and the decomposition result of the phase term on k-resolution Bessel basis is deduced. The m-D frequency is coarsely estimated by the one-to-one relationship between frequencies and basis functions. Then an algorithm is introduced to reduce the error and thus a finer estimation is obtained. Finally, the m-D frequency of each target is extracted by discrete echoes without phase shift ambiguity. Simulation experiments validate the effectiveness, and show that the proposed method outperforms the Sinusoidal Frequency Modulation Fourier Transform (SFMFT)-based method and Average Magnitude Difference Function (AMDF)-based method in estimation precision.
Micro-Doppler (m-D) feature extraction is significant for group target discrimination, while the methods for single target are invalid. An m-D feature extraction method of group targets is proposed based on signal orthogonal decomposition. First, the Sinusoidal Frequency-Modulated (SFM) form of m-D signals and the decomposition result of the phase term on k-resolution Bessel basis is deduced. The m-D frequency is coarsely estimated by the one-to-one relationship between frequencies and basis functions. Then an algorithm is introduced to reduce the error and thus a finer estimation is obtained. Finally, the m-D frequency of each target is extracted by discrete echoes without phase shift ambiguity. Simulation experiments validate the effectiveness, and show that the proposed method outperforms the Sinusoidal Frequency Modulation Fourier Transform (SFMFT)-based method and Average Magnitude Difference Function (AMDF)-based method in estimation precision.
2016, 38(12): 3063-3069.
doi: 10.11999/JEIT160008
Abstract:
In order to solve the problem of performance degradation when radar system is influenced by clutter from mainlobe and sidelobe, MIMO radar waveform design algorithm based on knowledge of range-spread target and clutter is investigated. Firstly, an optimization cost function is established, which includes mainlobe gain, sidelobe clutter suppression capability and Signal to Clutter plus Noise Ratio (SCNR) improvement. Secondly, to tackle the optimization problem, a relaxation is made to decouple spatial and temporal domain of the waveform matrix, beamforming and waveform design can be solved independently. Thirdly, L-BFGS algorithm is used to design the unimodular waveform matrix, beampattern with lower sidelobe and deep null is got. Based on maximization of SCNR, transmitted waveform and receiving filter are designed by iterative algorithm. Finally, the effectiveness of the proposed algorithm is verified by electromagnetic simulation of range-spread target.
In order to solve the problem of performance degradation when radar system is influenced by clutter from mainlobe and sidelobe, MIMO radar waveform design algorithm based on knowledge of range-spread target and clutter is investigated. Firstly, an optimization cost function is established, which includes mainlobe gain, sidelobe clutter suppression capability and Signal to Clutter plus Noise Ratio (SCNR) improvement. Secondly, to tackle the optimization problem, a relaxation is made to decouple spatial and temporal domain of the waveform matrix, beamforming and waveform design can be solved independently. Thirdly, L-BFGS algorithm is used to design the unimodular waveform matrix, beampattern with lower sidelobe and deep null is got. Based on maximization of SCNR, transmitted waveform and receiving filter are designed by iterative algorithm. Finally, the effectiveness of the proposed algorithm is verified by electromagnetic simulation of range-spread target.
2016, 38(12): 3070-3077.
doi: 10.11999/JEIT160667
Abstract:
In most of the frequency diverse array radars, a small linear frequency offset is always applied across the horizontal array elements, generating a different beampattern with coupling between the azimuth dimension and range dimension. Moreover, the grating lobes also exist in the range dimension. To overcome this, a method of beamforming with the vertical frequency diverse array is proposed. Based on the radar signal model, the characteristics of the time dependence and grating lobes in the transmit pattern are analyzed. Then the transmit-receive pattern with no grating lobes is synthesized by designing the parameters of the receive elements. Simulations results show the validity of the proposed method. The synthesized pattern is range-dependent only, and also has the smaller beamwidth of the mainlobe.
In most of the frequency diverse array radars, a small linear frequency offset is always applied across the horizontal array elements, generating a different beampattern with coupling between the azimuth dimension and range dimension. Moreover, the grating lobes also exist in the range dimension. To overcome this, a method of beamforming with the vertical frequency diverse array is proposed. Based on the radar signal model, the characteristics of the time dependence and grating lobes in the transmit pattern are analyzed. Then the transmit-receive pattern with no grating lobes is synthesized by designing the parameters of the receive elements. Simulations results show the validity of the proposed method. The synthesized pattern is range-dependent only, and also has the smaller beamwidth of the mainlobe.
2016, 38(12): 3078-3084.
doi: 10.11999/JEIT160995
Abstract:
A novel blind beamforming algorithm based on sparse Time-Frequency Decomposition (TFD) is proposed to solve the problems of existing blind beamforming algorithms: poor universality and the requirement of large amount of sampling data. In the proposed algorithm, the traditional Short-Time Fourier Transform (STFT) is first formulated as a sparse reconstruction problem. Then, a fast and efficient algorithm based on the alternating split Bregman technique is utilized to carry out the optimization. By combining the clustering and uncertainty set methods, the sparse-TFD results of the receiving data at each sensor are used to realize the estimation of Steering Vectors (SV). Finally, the optimal weight coefficients are achieved by substituting the estimated SV into the MVDR beamformer. The proposed algorithm hardly needs any specific statistical property of the receiving signals. Simulation results demonstrate that this algorithm can achieve superior output performance over the existing blind beamforming methods. It needs few snapshots with lower computational cost and has fast convergence rate, which makes the algorithm easy to utilize in practical applications.
A novel blind beamforming algorithm based on sparse Time-Frequency Decomposition (TFD) is proposed to solve the problems of existing blind beamforming algorithms: poor universality and the requirement of large amount of sampling data. In the proposed algorithm, the traditional Short-Time Fourier Transform (STFT) is first formulated as a sparse reconstruction problem. Then, a fast and efficient algorithm based on the alternating split Bregman technique is utilized to carry out the optimization. By combining the clustering and uncertainty set methods, the sparse-TFD results of the receiving data at each sensor are used to realize the estimation of Steering Vectors (SV). Finally, the optimal weight coefficients are achieved by substituting the estimated SV into the MVDR beamformer. The proposed algorithm hardly needs any specific statistical property of the receiving signals. Simulation results demonstrate that this algorithm can achieve superior output performance over the existing blind beamforming methods. It needs few snapshots with lower computational cost and has fast convergence rate, which makes the algorithm easy to utilize in practical applications.
2016, 38(12): 3085-3092.
doi: 10.11999/JEIT161013
Abstract:
In order to effectively suppress the noise of InSAR phase images and preserve the detailed fringe information, an adaptive phase filtering method based on local slope compensation and the Anisotropic Gaussian Filter (AGF) is proposed. Firstly, the topography-induced phase is approximately measured by local frequency estimation and removed from the original phase to eliminate the effect of the terrain topography. Secondly, the AGF with adaptive scale and orientation is developed to directionally filter out the noisy phase for the pixels with more homogeneous phase values. The scale of the AGF varies adaptively with the local coherence: a large-scaled AGF can better smooth the noise of low coherence areas, whereas a small-scaled AGF can better preserve the phase details of high coherence areas. Moreover, the orientation angle of the AGF is fast determined to select the identically distributed samples according to the maximum weighted coherent summation principle. The experimental results obtained via simulated and real data show that compared with commonly used filters, the proposed method achieves better performance in terms of residue reduction and fringe preservation.
In order to effectively suppress the noise of InSAR phase images and preserve the detailed fringe information, an adaptive phase filtering method based on local slope compensation and the Anisotropic Gaussian Filter (AGF) is proposed. Firstly, the topography-induced phase is approximately measured by local frequency estimation and removed from the original phase to eliminate the effect of the terrain topography. Secondly, the AGF with adaptive scale and orientation is developed to directionally filter out the noisy phase for the pixels with more homogeneous phase values. The scale of the AGF varies adaptively with the local coherence: a large-scaled AGF can better smooth the noise of low coherence areas, whereas a small-scaled AGF can better preserve the phase details of high coherence areas. Moreover, the orientation angle of the AGF is fast determined to select the identically distributed samples according to the maximum weighted coherent summation principle. The experimental results obtained via simulated and real data show that compared with commonly used filters, the proposed method achieves better performance in terms of residue reduction and fringe preservation.
2016, 38(12): 3093-3099.
doi: 10.11999/JEIT161035
Abstract:
This paper studies on the feature extraction methods for the classification of helicopter, propeller-driven aircraft, and turbojet using a conventional narrow-band radar system. In the modern battlefield, the helicopter, propeller aircraft and jet aircraft with different motor performances each bear an important task. But the classification performance of the traditional features for the three types of aircraft target classification is not good enough, so the Fractional Fourier Transform (FrFT) is introduced. Based on the existing feature extraction method, the fractional order features of three kinds of aircraft targets are extracted from the fractional domain after FrFT to extend feature domain. Then, the effective features are selected from all extracted features and the classification of the three categories via linear Relevance Vector Machine (RVM) is realized. The experiments demonstrate that the proposed fractional features can improve the classification performance in comparison with some existing features from the time-domain and Doppler-frequency domain.
This paper studies on the feature extraction methods for the classification of helicopter, propeller-driven aircraft, and turbojet using a conventional narrow-band radar system. In the modern battlefield, the helicopter, propeller aircraft and jet aircraft with different motor performances each bear an important task. But the classification performance of the traditional features for the three types of aircraft target classification is not good enough, so the Fractional Fourier Transform (FrFT) is introduced. Based on the existing feature extraction method, the fractional order features of three kinds of aircraft targets are extracted from the fractional domain after FrFT to extend feature domain. Then, the effective features are selected from all extracted features and the classification of the three categories via linear Relevance Vector Machine (RVM) is realized. The experiments demonstrate that the proposed fractional features can improve the classification performance in comparison with some existing features from the time-domain and Doppler-frequency domain.
2016, 38(12): 3100-3106.
doi: 10.11999/JEIT160882
Abstract:
Coherent sources commonly exist in scenarios with multipath effect. How to decorrelate coherent sources is traditionally a problem urgently to be solved in the array signal processing domain. Especially for spatially adjacent coherent sources, the performance of the estimation of Direction Of Arrival (DOA) remains to be improved. A DOA estimation method based on spatial filtering is proposed for spatially adjacent coherent sources. Multiple coherent sources are separated by spatial filtering and the DOAs are estimated respectively afterwards. The performance of the DOA estimation is enhanced by refining the filter parameters and the DOAs of the coherent sources iteratively. To extend its application to non-uniform linear array, the virtual array technique is adopted. The computer simulation results indicate that the proposed algorithm has better DOA estimation performance than the existing methods. In the scenario of sufficiently high Signal to Noise Ratio (SNR), the Root Mean Square Error (RMSE) could achieve Cramer-Rao Bound (CRB). The effectiveness and the superiority of the proposed method for spatially adjacent coherent sources are validated by the simulation results.
Coherent sources commonly exist in scenarios with multipath effect. How to decorrelate coherent sources is traditionally a problem urgently to be solved in the array signal processing domain. Especially for spatially adjacent coherent sources, the performance of the estimation of Direction Of Arrival (DOA) remains to be improved. A DOA estimation method based on spatial filtering is proposed for spatially adjacent coherent sources. Multiple coherent sources are separated by spatial filtering and the DOAs are estimated respectively afterwards. The performance of the DOA estimation is enhanced by refining the filter parameters and the DOAs of the coherent sources iteratively. To extend its application to non-uniform linear array, the virtual array technique is adopted. The computer simulation results indicate that the proposed algorithm has better DOA estimation performance than the existing methods. In the scenario of sufficiently high Signal to Noise Ratio (SNR), the Root Mean Square Error (RMSE) could achieve Cramer-Rao Bound (CRB). The effectiveness and the superiority of the proposed method for spatially adjacent coherent sources are validated by the simulation results.
2016, 38(12): 3107-3113.
doi: 10.11999/JEIT160873
Abstract:
A fast iterative optimization algorithm for sum and difference patterns with controllable accuracy and range of angle measurement is proposed, which can be used in fast targets location under wide angle coverage and accurate angle tracking of maneuvering targets. Under expected accuracy or range of angle estimation, the expected main-lobe of sum and difference patterns are modified in the form of sinc function and its derivative respectively, then the sum and difference patterns are rapidly synthesized with the proposed regional weighting pattern synthesis algorithm in the close form. The proposed iterative optimization algorithm can effectively maximize the range of effective angle measurement at a given input signal to noise ratio and angular accuracy, or get optimal accuracy within a given range of angle measurement.
A fast iterative optimization algorithm for sum and difference patterns with controllable accuracy and range of angle measurement is proposed, which can be used in fast targets location under wide angle coverage and accurate angle tracking of maneuvering targets. Under expected accuracy or range of angle estimation, the expected main-lobe of sum and difference patterns are modified in the form of sinc function and its derivative respectively, then the sum and difference patterns are rapidly synthesized with the proposed regional weighting pattern synthesis algorithm in the close form. The proposed iterative optimization algorithm can effectively maximize the range of effective angle measurement at a given input signal to noise ratio and angular accuracy, or get optimal accuracy within a given range of angle measurement.
2016, 38(12): 3114-3121.
doi: 10.11999/JEIT160372
Abstract:
Extended targets usually generate multiple measurements in high resolution radar systems. Existing algorithms of the Random Finite Set (RFS) assume that the measurement number of extended targets follows Poisson distribution in a general way. However, this assumption is inconsistent with actual situations. Considering this issue, a Multi-Bernoulli Extended Target Probability Hypothesis Density (MB-ET-PHD) tracking method is proposed. First, this method assumes that the measurement number of extended targets is Multi-Bernoulli (MB) distributed. Then, its update equation is derived by using the FInite Set STatistics (FISST) multi-target calculus. Finally, simulated results of Gaussian Mixture (GM) framework are given. The simulation results show that the proposed method can obtain better tracking performance compared with the Poisson ET-PHD method.
Extended targets usually generate multiple measurements in high resolution radar systems. Existing algorithms of the Random Finite Set (RFS) assume that the measurement number of extended targets follows Poisson distribution in a general way. However, this assumption is inconsistent with actual situations. Considering this issue, a Multi-Bernoulli Extended Target Probability Hypothesis Density (MB-ET-PHD) tracking method is proposed. First, this method assumes that the measurement number of extended targets is Multi-Bernoulli (MB) distributed. Then, its update equation is derived by using the FInite Set STatistics (FISST) multi-target calculus. Finally, simulated results of Gaussian Mixture (GM) framework are given. The simulation results show that the proposed method can obtain better tracking performance compared with the Poisson ET-PHD method.
2016, 38(12): 3122-3128.
doi: 10.11999/JEIT160812
Abstract:
In highly squinted and high resolution SAR, the data receiving range bin can be continually adjusted to minimize the sampled data in echo collection, but the staggered range cell migration needs extra processing to be eliminated. To enlarge the imaging scene size with a relatively high resolution, sliding spotlight working mode could be used, yet this mode may cause azimuth ambiguity. In this paper, the key technique of highly squinted sliding spotlight SAR imaging with varied receiving range bin is investigated, a sub-aperture based upsampling ambiguity-resolving method is utilized and a new beam segmentation based 2-step PFA method is proposed, which incorporates the stagger compensation to the procedure of motion compensation. The method performs beam segmentation via digital spotlighting preprocessing firstly to generate multiple full-resolution phase histories of smaller image patches, which allow the approximation of planar wavefront in traditional PFA. PFA is used to produce focused fine resolution image for each small patch．Finally, all focused image patches are seamed together to get a full image. This divide-and-conquer approach breaks the image size limit in traditional PFA, and extensively enlarges the valid focused scene suitable for sliding spotlight mode. This new algorithm is validated to be effective and efficient via real data experiments．
In highly squinted and high resolution SAR, the data receiving range bin can be continually adjusted to minimize the sampled data in echo collection, but the staggered range cell migration needs extra processing to be eliminated. To enlarge the imaging scene size with a relatively high resolution, sliding spotlight working mode could be used, yet this mode may cause azimuth ambiguity. In this paper, the key technique of highly squinted sliding spotlight SAR imaging with varied receiving range bin is investigated, a sub-aperture based upsampling ambiguity-resolving method is utilized and a new beam segmentation based 2-step PFA method is proposed, which incorporates the stagger compensation to the procedure of motion compensation. The method performs beam segmentation via digital spotlighting preprocessing firstly to generate multiple full-resolution phase histories of smaller image patches, which allow the approximation of planar wavefront in traditional PFA. PFA is used to produce focused fine resolution image for each small patch．Finally, all focused image patches are seamed together to get a full image. This divide-and-conquer approach breaks the image size limit in traditional PFA, and extensively enlarges the valid focused scene suitable for sliding spotlight mode. This new algorithm is validated to be effective and efficient via real data experiments．
2016, 38(12): 3129-3136.
doi: 10.11999/JEIT160177
Abstract:
In order to make full use of the joint sparse physical characteristics of the radar echo to improve imaging performance. A novel super resolution Inverse SAR (ISAR) imaging method based on distributed compressed sensing theory is proposed. Firstly, the joint sparse echo model of the random chirp frequency-stepped signal is built and the pulse compression processing of each sub-pulse is processed. Secondly, owing to different random patterns of each group, different measurement matrices are constructed in accordance with the random pattern of sub-pulse signal. Then the corresponding compressed sensing model of the echo is built and the supper resolution range profile is obtained via the distributed compressed sensing theory. Finally, the supper resolution inverse synthetic aperture radar image can be obtained by a fast compressed sensing reconstruction algorithm, which is used to achieve the high resolution reconstruction in azimuth direction based on the sparse features. Theoretical analysis and simulation results show that the proposed method has the characteristics of high reconstruction accuracy, low sampling rate and strong anti-noise performance.
In order to make full use of the joint sparse physical characteristics of the radar echo to improve imaging performance. A novel super resolution Inverse SAR (ISAR) imaging method based on distributed compressed sensing theory is proposed. Firstly, the joint sparse echo model of the random chirp frequency-stepped signal is built and the pulse compression processing of each sub-pulse is processed. Secondly, owing to different random patterns of each group, different measurement matrices are constructed in accordance with the random pattern of sub-pulse signal. Then the corresponding compressed sensing model of the echo is built and the supper resolution range profile is obtained via the distributed compressed sensing theory. Finally, the supper resolution inverse synthetic aperture radar image can be obtained by a fast compressed sensing reconstruction algorithm, which is used to achieve the high resolution reconstruction in azimuth direction based on the sparse features. Theoretical analysis and simulation results show that the proposed method has the characteristics of high reconstruction accuracy, low sampling rate and strong anti-noise performance.
2016, 38(12): 3137-3143.
doi: 10.11999/JEIT160784
Abstract:
Due to the complexity of SAR imaging of maneuvering platform, traditional imaging algorithms are unavailable in the diving stage with high squint angle. To deal with these problems, the geometric model in diving stage of high squint SAR is constructed and discussed in detail. Furthermore, the spectrum support zone is analyzed. Then, a wavenumber-domain imaging algorithm for high squint diving SAR based on axes rotation is proposed. The wave-number domain imaging processing for sub-aperture data obtains high-usage of spectrum by axes rotation to guarantee the resolution. By doing so, zero-pudding operations in imaging domain are decreased effectively, which lead to the promotion of efficiency. Simulation results and real data processing are presented to valid the superiority of the proposed approach.
Due to the complexity of SAR imaging of maneuvering platform, traditional imaging algorithms are unavailable in the diving stage with high squint angle. To deal with these problems, the geometric model in diving stage of high squint SAR is constructed and discussed in detail. Furthermore, the spectrum support zone is analyzed. Then, a wavenumber-domain imaging algorithm for high squint diving SAR based on axes rotation is proposed. The wave-number domain imaging processing for sub-aperture data obtains high-usage of spectrum by axes rotation to guarantee the resolution. By doing so, zero-pudding operations in imaging domain are decreased effectively, which lead to the promotion of efficiency. Simulation results and real data processing are presented to valid the superiority of the proposed approach.
2016, 38(12): 3144-3151.
doi: 10.11999/JEIT161025
Abstract:
Three dimensional imaging of space micro-motion target has significant advantages on target information awareness, which is crucial to effectively realize space target imaging, classification and recognition. In this paper, through the L type antenna array imaging system, an interferometric three dimensional imaging method for space micro-motion target is proposed based on the improved Particle Swarm Optimization (PSO) algorithm. Firstly, the Doppler effect in the received signal is analyzed, and the corresponding parametric model is established. Then, the Doppler phase term of the received signal is reconstructed by using the proposed optimization method. Through interferometric processing and analyzing the quantitative relationship between interferometric phase difference and real coordinate, the three dimensional coordinates and image can be obtained. Compared with the existing methods, the proposed method can reconstruct the real coordinates and three dimensional image of micro-motion target with and without occlusion effect. It also has good robustness. Finally, simulations validate the effectiveness of the proposed method.
Three dimensional imaging of space micro-motion target has significant advantages on target information awareness, which is crucial to effectively realize space target imaging, classification and recognition. In this paper, through the L type antenna array imaging system, an interferometric three dimensional imaging method for space micro-motion target is proposed based on the improved Particle Swarm Optimization (PSO) algorithm. Firstly, the Doppler effect in the received signal is analyzed, and the corresponding parametric model is established. Then, the Doppler phase term of the received signal is reconstructed by using the proposed optimization method. Through interferometric processing and analyzing the quantitative relationship between interferometric phase difference and real coordinate, the three dimensional coordinates and image can be obtained. Compared with the existing methods, the proposed method can reconstruct the real coordinates and three dimensional image of micro-motion target with and without occlusion effect. It also has good robustness. Finally, simulations validate the effectiveness of the proposed method.
2016, 38(12): 3152-3158.
doi: 10.11999/JEIT160656
Abstract:
The GEO SAR has its own features such as wide coverage and short revisit time. However, when the GEO SAR is both used as a transmitter and a receiver, its advantages is not well exploited. If an airplane or a LEO satellite is adopted as a platform of the receiver, not only the interesting regions can be observed flexibly, but also finer resolution can be achieved. However, the geometry of the BiSAR is complicated, so it is not easy to acquire how much resolution an arbitrary BiSAR system can reach. Thus starting with the resolution on the basic plane of a BiSAR system, and combined with the resolutions projection relation between the basic plane and the plane tangent to the earths surface, the resolution shapes expression on the ground can be got finally. Based on the expression, the resolution of a BiSAR system can be assessed, and finer resolution can be realized through optimizing two parameters, including signal bandwidth and synthetic aperture time. Finally, the simulation results validate the effectiveness of the proposed method.
The GEO SAR has its own features such as wide coverage and short revisit time. However, when the GEO SAR is both used as a transmitter and a receiver, its advantages is not well exploited. If an airplane or a LEO satellite is adopted as a platform of the receiver, not only the interesting regions can be observed flexibly, but also finer resolution can be achieved. However, the geometry of the BiSAR is complicated, so it is not easy to acquire how much resolution an arbitrary BiSAR system can reach. Thus starting with the resolution on the basic plane of a BiSAR system, and combined with the resolutions projection relation between the basic plane and the plane tangent to the earths surface, the resolution shapes expression on the ground can be got finally. Based on the expression, the resolution of a BiSAR system can be assessed, and finer resolution can be realized through optimizing two parameters, including signal bandwidth and synthetic aperture time. Finally, the simulation results validate the effectiveness of the proposed method.
2016, 38(12): 3159-3165.
doi: 10.11999/JEIT161006
Abstract:
For missile-borne forward-looking SAR in curve trajectory, time-variant motion errors are ignored, leading to degradations in imaging result. This paper proposes a missile-borne forward-looking SAR algorithm based on motion compensation to solve the above problem. The acceleration is divided into forward-looking acceleration and cross-track acceleration, the second phase error and cubic phase error caused by these two accelerations are analyzed in detail. For the cross-track acceleration component, it can be divided into the acceleration vertical to the imaging plane and the one in imaging plane further. Via such dividing, the phase errors caused by acceleration are compensated by using vectorial methods. Moreover, for range migration, it is compensated through Nonlinear Chirp Scaling (NCS) approach based on the accurate 2-D spectrum acquired by the Method of Series Reversion (MSR). The simulated results are given to illustrate the validity of the proposed algorithm.
For missile-borne forward-looking SAR in curve trajectory, time-variant motion errors are ignored, leading to degradations in imaging result. This paper proposes a missile-borne forward-looking SAR algorithm based on motion compensation to solve the above problem. The acceleration is divided into forward-looking acceleration and cross-track acceleration, the second phase error and cubic phase error caused by these two accelerations are analyzed in detail. For the cross-track acceleration component, it can be divided into the acceleration vertical to the imaging plane and the one in imaging plane further. Via such dividing, the phase errors caused by acceleration are compensated by using vectorial methods. Moreover, for range migration, it is compensated through Nonlinear Chirp Scaling (NCS) approach based on the accurate 2-D spectrum acquired by the Method of Series Reversion (MSR). The simulated results are given to illustrate the validity of the proposed algorithm.
2016, 38(12): 3166-3173.
doi: 10.11999/JEIT160785
Abstract:
Due to the acceleration in the moving of the platform, the SAR imaging of highly squinted maneuvering platform becomes a problem to be solved urgently. The majority of currently existing range models fail in taking the azimuth-variation into consideration, so the well-focused images are not obtained. To solve these problems, a new azimuth-variation range model under the non-linear trajectory caused by acceleration is constructed. Based on this model, a wavenumber-domain imaging algorithm for high squint SAR is proposed, which takes advantage of the azimuth-variation filter to remove the azimuth chirp rate changes and Doppler center changes caused by the acceleration. By the analysis of the model error, the precision of the model is confirmed. Contrast simulation results are presented to valid the superiority of the proposed approach.
Due to the acceleration in the moving of the platform, the SAR imaging of highly squinted maneuvering platform becomes a problem to be solved urgently. The majority of currently existing range models fail in taking the azimuth-variation into consideration, so the well-focused images are not obtained. To solve these problems, a new azimuth-variation range model under the non-linear trajectory caused by acceleration is constructed. Based on this model, a wavenumber-domain imaging algorithm for high squint SAR is proposed, which takes advantage of the azimuth-variation filter to remove the azimuth chirp rate changes and Doppler center changes caused by the acceleration. By the analysis of the model error, the precision of the model is confirmed. Contrast simulation results are presented to valid the superiority of the proposed approach.
2016, 38(12): 3174-3181.
doi: 10.11999/JEIT161016
Abstract:
In One-Stationary Bistatic Synthetic Aperture Radar (OS-BiSAR) imaging, imprecise description of 2-D range-azimuth space-variant property usually leads to deterioration of final SAR image rapidly. In order to solve this issue, a new ellipse model is proposed to precisely describe range-azimuth space-variant property of OS-BiSAR with large baseline, and an improved Non-Linear Chirp Scaling (NLCS) algorithm is also derived based on this model. First, a phase de-ramp operation is performed to remove the linear Range Cell Migration (RCM) and Doppler centroid in range frequency domain. Then, the residual RCM and high order range-azimuth coupling terms are removed. Thirdly, a new ellipse model is established to describe range-azimuth space-variant property of OS-BiSAR, and then the azimuth frequency modulation rate of space-variant echo is analyzed. Moreover, azimuth scaling function of NLCS and azimuth compression factors are re-derived. Theoretical analysis and simulation results show that the proposed model not only reveals the property of 2-D azimuth-variant in OS-BiSAR, but also provides a precise analytical expression to depict the 2-D range-azimuth space-variant property of OS-BiSAR. Furthermore, simulation results validate that the improved NLCS algorithm based on this new model has high imaging performance.
In One-Stationary Bistatic Synthetic Aperture Radar (OS-BiSAR) imaging, imprecise description of 2-D range-azimuth space-variant property usually leads to deterioration of final SAR image rapidly. In order to solve this issue, a new ellipse model is proposed to precisely describe range-azimuth space-variant property of OS-BiSAR with large baseline, and an improved Non-Linear Chirp Scaling (NLCS) algorithm is also derived based on this model. First, a phase de-ramp operation is performed to remove the linear Range Cell Migration (RCM) and Doppler centroid in range frequency domain. Then, the residual RCM and high order range-azimuth coupling terms are removed. Thirdly, a new ellipse model is established to describe range-azimuth space-variant property of OS-BiSAR, and then the azimuth frequency modulation rate of space-variant echo is analyzed. Moreover, azimuth scaling function of NLCS and azimuth compression factors are re-derived. Theoretical analysis and simulation results show that the proposed model not only reveals the property of 2-D azimuth-variant in OS-BiSAR, but also provides a precise analytical expression to depict the 2-D range-azimuth space-variant property of OS-BiSAR. Furthermore, simulation results validate that the improved NLCS algorithm based on this new model has high imaging performance.
2016, 38(12): 3182-3188.
doi: 10.11999/JEIT160603
Abstract:
A new method for attitude estimation for space satellite targets is presented by extracting typical linear structures in ISAR imaging sequence and using information of targets position in orbits to analyze the three-dimensional attitude of space satellite targets. With the analyzing process for space satellite targets geometric structures, the algorithm utilizes Radon transformation to realize the extraction of linear structures, like solar wings, planar antennas, in ISAR imaging sequences. After finishing the relevance of these linear structures among different frames, the angle information of typical linear structures in range-Doppler plane is extracted. At last, with targets position information in orbits, a matrix sequence of the ISAR range-Doppler projection is acquired to estimate the three-dimensional attitude of linear structures, and realize exact solution of space satellite targets attitude. The simulation experiment result illustrates that the algorithm can realize the attitude estimation of typical units in space satellite targets, and the multistatic model algorithm shows its advantage in estimation accuracy.
A new method for attitude estimation for space satellite targets is presented by extracting typical linear structures in ISAR imaging sequence and using information of targets position in orbits to analyze the three-dimensional attitude of space satellite targets. With the analyzing process for space satellite targets geometric structures, the algorithm utilizes Radon transformation to realize the extraction of linear structures, like solar wings, planar antennas, in ISAR imaging sequences. After finishing the relevance of these linear structures among different frames, the angle information of typical linear structures in range-Doppler plane is extracted. At last, with targets position information in orbits, a matrix sequence of the ISAR range-Doppler projection is acquired to estimate the three-dimensional attitude of linear structures, and realize exact solution of space satellite targets attitude. The simulation experiment result illustrates that the algorithm can realize the attitude estimation of typical units in space satellite targets, and the multistatic model algorithm shows its advantage in estimation accuracy.
2016, 38(12): 3189-3196.
doi: 10.11999/JEIT160022
Abstract:
By employing the Maximum Correntropy Criterion (MCC) based cost function in PARAllel FACtor (PARAFAC), the MCC-PARAFAC algorithm is deduced, which can be utilized for the parallel factor under impulsive noise environments. The MCC-PARAFAC algorithm is applied to parameter estimation in bistatic MIMO radar under impulsive noise environment. The proposed method can suppress the impulse noise interference and has better estimation performance. Furthermore, the estimated parameters are automatically paired without the additional pairing method. Simulation results verify the effectiveness of the proposed method.
By employing the Maximum Correntropy Criterion (MCC) based cost function in PARAllel FACtor (PARAFAC), the MCC-PARAFAC algorithm is deduced, which can be utilized for the parallel factor under impulsive noise environments. The MCC-PARAFAC algorithm is applied to parameter estimation in bistatic MIMO radar under impulsive noise environment. The proposed method can suppress the impulse noise interference and has better estimation performance. Furthermore, the estimated parameters are automatically paired without the additional pairing method. Simulation results verify the effectiveness of the proposed method.
2016, 38(12): 3197-3204.
doi: 10.11999/JEIT160041
Abstract:
Ionospheric phase decontamination is a key technology in signal processing of sky-wave Over-The- Horizon Radar (OTHR). Due to the inaccuracy of the models and the complexity of the ionosphere, the accuracy of the existing algorithms is not satisfactory when the phase changes too fast. A new ionospheric phase decontamination algorithm is proposed based on the Maximum-Likelihood (ML) method. In this algorithm, the signal is modeled as a phase polynomial, and estimation of the perturbation phase is achieved by maximizing the likelihood function. To avoid matrix inversion in the ML method, the ML issue is further transformed to a least-squares issue. The coefficients of phase are solved by the genetic algorithm. The simulation results show that, compared with the traditional methods, the proposed algorithm has the following advantages: compared with the HRR algorithm and the CED algorithm, the algorithm proposed in this paper has higher accuracy, and the signal spectrum after decontamination is more sharp. Under the situation of serious phase contamination, the proposed algorithm still has higher precision, accordingly, the proposed algorithm is more advantageous to extract the target information. This algorithm adopts higher-order polynomials, which avoids segmented processing and computing the inverse of matrix, thus the computation process is simplified.
Ionospheric phase decontamination is a key technology in signal processing of sky-wave Over-The- Horizon Radar (OTHR). Due to the inaccuracy of the models and the complexity of the ionosphere, the accuracy of the existing algorithms is not satisfactory when the phase changes too fast. A new ionospheric phase decontamination algorithm is proposed based on the Maximum-Likelihood (ML) method. In this algorithm, the signal is modeled as a phase polynomial, and estimation of the perturbation phase is achieved by maximizing the likelihood function. To avoid matrix inversion in the ML method, the ML issue is further transformed to a least-squares issue. The coefficients of phase are solved by the genetic algorithm. The simulation results show that, compared with the traditional methods, the proposed algorithm has the following advantages: compared with the HRR algorithm and the CED algorithm, the algorithm proposed in this paper has higher accuracy, and the signal spectrum after decontamination is more sharp. Under the situation of serious phase contamination, the proposed algorithm still has higher precision, accordingly, the proposed algorithm is more advantageous to extract the target information. This algorithm adopts higher-order polynomials, which avoids segmented processing and computing the inverse of matrix, thus the computation process is simplified.
2016, 38(12): 3205-3211.
doi: 10.11999/JEIT161075
Abstract:
VHF radar has unique advantage in anti-stealth and resisting anti-radiation missile. It plays an important role in modern antiaircraft system. However, the multipath signal often brings difficulties to the altitude measurement of VHF radar. Combining with the characteristics of VHF array radar and array multipath signal model, this paper summarizes and concludes three VHF radar height measurement methods based on the traditional Maximum Likelihood (ML) algorithm: the altitude measurement method based on the temporal-spatial sequential ML algorithm; the altitude measurement method based on the improved temporal-spatial sequential ML algorithm; the altitude measurement method based on the Refined Maximum Likelihood (RML) algorithm. This paper presents the theoretical performance analysis of these methods, the relationship between three methods, and the results of computer simulation experiments. Finally some meaningful conclusions are given.
VHF radar has unique advantage in anti-stealth and resisting anti-radiation missile. It plays an important role in modern antiaircraft system. However, the multipath signal often brings difficulties to the altitude measurement of VHF radar. Combining with the characteristics of VHF array radar and array multipath signal model, this paper summarizes and concludes three VHF radar height measurement methods based on the traditional Maximum Likelihood (ML) algorithm: the altitude measurement method based on the temporal-spatial sequential ML algorithm; the altitude measurement method based on the improved temporal-spatial sequential ML algorithm; the altitude measurement method based on the Refined Maximum Likelihood (RML) algorithm. This paper presents the theoretical performance analysis of these methods, the relationship between three methods, and the results of computer simulation experiments. Finally some meaningful conclusions are given.
2016, 38(12): 3212-3218.
doi: 10.11999/JEIT160002
Abstract:
For the problem of signal interception and fast moving jammers in MIMO radar application, this paper designs a new transmit beampattern through optimizing the transmit beamforming matrix based on the minimization of integrated-sidelobe. This new transmit beampattern can not only focus the transmit energy on the desired spatial sector but decrease the level of sidelobe as well as form a nulling at the jammers direction with the prior knowledge. At the receiving, a new robust beamforming method based on non-uniform generalized diagonal loading is proposed, which can strengthen the robustness of receiving beamformer against different errors and broaden the nulling adaptively. Through the optimization of joint transmit and receive beampattern for MIMO radar, the interception of signal and fast moving jamming are addressed, which improve the performance of MIMO radar from both transmitting and receiving. Simulation results and comparisons with existing methods demonstrate the feasibility and effectiveness of the proposed methods.
For the problem of signal interception and fast moving jammers in MIMO radar application, this paper designs a new transmit beampattern through optimizing the transmit beamforming matrix based on the minimization of integrated-sidelobe. This new transmit beampattern can not only focus the transmit energy on the desired spatial sector but decrease the level of sidelobe as well as form a nulling at the jammers direction with the prior knowledge. At the receiving, a new robust beamforming method based on non-uniform generalized diagonal loading is proposed, which can strengthen the robustness of receiving beamformer against different errors and broaden the nulling adaptively. Through the optimization of joint transmit and receive beampattern for MIMO radar, the interception of signal and fast moving jamming are addressed, which improve the performance of MIMO radar from both transmitting and receiving. Simulation results and comparisons with existing methods demonstrate the feasibility and effectiveness of the proposed methods.
2016, 38(12): 3219-3223.
doi: 10.11999/JEIT160981
Abstract:
To meet the need of the real application, this paper proposes a power allocation algorithm for multiple target localization, which tries to get the quick optimal allocation of the limited power resources in the MIMO radar. Firstly, Cramr-Rao Lower Bound (CRLB) of the Mean Square Error (MSE) of the multi-target localization is given, and CRLB is used as a cost function to allocate the power resource. Then, an Alternating Global Optimal Algorithm (AGOA) is designed which can be used in power allocation of multi-target localization, the related Pareto sets to achieve the fast allocation of the power resources. Finally, the simulation results show that the AGOA can quickly achieve the optimal allocation of the limited power allocation in MIMO radar, and can significantly enhance the precision of the multiple target localization.
To meet the need of the real application, this paper proposes a power allocation algorithm for multiple target localization, which tries to get the quick optimal allocation of the limited power resources in the MIMO radar. Firstly, Cramr-Rao Lower Bound (CRLB) of the Mean Square Error (MSE) of the multi-target localization is given, and CRLB is used as a cost function to allocate the power resource. Then, an Alternating Global Optimal Algorithm (AGOA) is designed which can be used in power allocation of multi-target localization, the related Pareto sets to achieve the fast allocation of the power resources. Finally, the simulation results show that the AGOA can quickly achieve the optimal allocation of the limited power allocation in MIMO radar, and can significantly enhance the precision of the multiple target localization.
2016, 38(12): 3224-3229.
doi: 10.11999/JEIT160132
Abstract:
Owing to the heavy spread of eigenspectrum of the population covariance matrix under finite training samples condition, it is a challenge to estimate the clutter Degrees of Freedom (DoF) in airborne forward-looking radar. In this work, a method for estimation the clutters DoF is proposed. In order to estimate the clutters DoF, an idea from sources detection by Minimum Description Length (MDL) criterion is borrowed, and the parametric probability model is formed based on the eigenvalues statistical distribution properties from Random Matrix Theory (RMT). The proposed method is effective to estimate the clutters DoF under finite training samples condition, and the simulation results verify the efficiency of the proposed method.
Owing to the heavy spread of eigenspectrum of the population covariance matrix under finite training samples condition, it is a challenge to estimate the clutter Degrees of Freedom (DoF) in airborne forward-looking radar. In this work, a method for estimation the clutters DoF is proposed. In order to estimate the clutters DoF, an idea from sources detection by Minimum Description Length (MDL) criterion is borrowed, and the parametric probability model is formed based on the eigenvalues statistical distribution properties from Random Matrix Theory (RMT). The proposed method is effective to estimate the clutters DoF under finite training samples condition, and the simulation results verify the efficiency of the proposed method.
2016, 38(12): 3230-3237.
doi: 10.11999/JEIT160954
Abstract:
Space time processing is an effective method for the suppression of clutters and the power integration of target echo for airborne passive radar. However, it needs long Coherent Processing Intervals (CPI) to improve target Signal-to-Noise Ratio (SNR) because of the weak target in passive radar, which leads to range migration and integration loss, and then lowers the system performance. Focusing on this problem, a range migration compensation algorithm is proposed, which combines Keystone transform with 3DT-SAP algorithm perfectly. This algorithm is efficient in computation and owns the potential for real time implementation. In addition, it can compensate the range migration with little power loss at the same time of clutter suppression. Simulations show that the proposed algorithm compensates the range migration of targets with different velocities and different powers effectively when suppressing clutters fully, which means it is an efficient and high-performance range migration compensation algorithm for airborne passive radar.
Space time processing is an effective method for the suppression of clutters and the power integration of target echo for airborne passive radar. However, it needs long Coherent Processing Intervals (CPI) to improve target Signal-to-Noise Ratio (SNR) because of the weak target in passive radar, which leads to range migration and integration loss, and then lowers the system performance. Focusing on this problem, a range migration compensation algorithm is proposed, which combines Keystone transform with 3DT-SAP algorithm perfectly. This algorithm is efficient in computation and owns the potential for real time implementation. In addition, it can compensate the range migration with little power loss at the same time of clutter suppression. Simulations show that the proposed algorithm compensates the range migration of targets with different velocities and different powers effectively when suppressing clutters fully, which means it is an efficient and high-performance range migration compensation algorithm for airborne passive radar.
2016, 38(12): 3238-3244.
doi: 10.11999/JEIT161000
Abstract:
To improve the detection accuracy of urban built-up areas from Polarimetric Synthetic Aperture Radar (PolSAR) images, this paper proposes a new built-up area detection method based on nonstationarity and polarization coherency coefficient ratio. Firstly, the PolSAR image is filtered and decomposed into several sub-aperture images along the azimuth direction. Then nonstationarity detection and polarization coherency coefficient ratio are combined to determine the class label of pixels. On the basis of the traditional nonstationarity detection method, this paper introduces a new polarization coherency coefficient ratio to remove the false alarms of natural areas and to improve the overall detection accuracy. Experimental results using spaceborne and airborne PolSAR data demonstrate the effectiveness of the proposed method.
To improve the detection accuracy of urban built-up areas from Polarimetric Synthetic Aperture Radar (PolSAR) images, this paper proposes a new built-up area detection method based on nonstationarity and polarization coherency coefficient ratio. Firstly, the PolSAR image is filtered and decomposed into several sub-aperture images along the azimuth direction. Then nonstationarity detection and polarization coherency coefficient ratio are combined to determine the class label of pixels. On the basis of the traditional nonstationarity detection method, this paper introduces a new polarization coherency coefficient ratio to remove the false alarms of natural areas and to improve the overall detection accuracy. Experimental results using spaceborne and airborne PolSAR data demonstrate the effectiveness of the proposed method.
2016, 38(12): 3245-3251.
doi: 10.11999/JEIT160021
Abstract:
High resolution airborne Interferometric Synthetic Aperture Radar (InSAR) is one of the important methods to generate high precision Digital Elevation Model (DEM). Due to the phase pattern difference between the two antennas, the interferometric phase bias varies with the range. The traditional interferometric calibration regards the phase bias as a constant, and it is not able to correct the range-variant phase errors. Therefore, there are range-variant errors of the reconstructed elevation. To solve this problem, this paper presents a calibration method that uses polynomial to fit the interference phase bias. At last, a set of real airborne InSAR data are used to validate the method, and the experimental results show that the proposed method can solve the problem of range-variant height errors in high resolution airborne InSAR effectively.
High resolution airborne Interferometric Synthetic Aperture Radar (InSAR) is one of the important methods to generate high precision Digital Elevation Model (DEM). Due to the phase pattern difference between the two antennas, the interferometric phase bias varies with the range. The traditional interferometric calibration regards the phase bias as a constant, and it is not able to correct the range-variant phase errors. Therefore, there are range-variant errors of the reconstructed elevation. To solve this problem, this paper presents a calibration method that uses polynomial to fit the interference phase bias. At last, a set of real airborne InSAR data are used to validate the method, and the experimental results show that the proposed method can solve the problem of range-variant height errors in high resolution airborne InSAR effectively.
2016, 38(12): 3252-3260.
doi: 10.11999/JEIT161031
Abstract:
Wind farms clutters have the characteristics of strong scattering and the Doppler spectrum spreading due to the blades rotation, the radar echoes can not be filtered out easily using the traditional ground clutter filter, hence causing the false detection and identification of meteorological targets in the process of target detection, which is an important influence factor on the new generation weather radar echoes. Based on the analysis of wind farms echoes characteristics distinguished from those of meteorological target echoes, some suitable feature parameters are chosen, and a robust good adaptive fuzzy logic system of wind farms clutters detection and identification is developed by using the secondary products (Level II) data and the Fuzzy Inference System (FIS), in which the membership functions of each feature parameters and the corresponding logical rules are defined by constructing probability distribution histogram and the one dimensional range distribution of the corresponding feature parameters. Several groups of typical Level II data are collected to test and verify the proposed method, the experimental results demonstrate the reliability of the proposed algorithm.
Wind farms clutters have the characteristics of strong scattering and the Doppler spectrum spreading due to the blades rotation, the radar echoes can not be filtered out easily using the traditional ground clutter filter, hence causing the false detection and identification of meteorological targets in the process of target detection, which is an important influence factor on the new generation weather radar echoes. Based on the analysis of wind farms echoes characteristics distinguished from those of meteorological target echoes, some suitable feature parameters are chosen, and a robust good adaptive fuzzy logic system of wind farms clutters detection and identification is developed by using the secondary products (Level II) data and the Fuzzy Inference System (FIS), in which the membership functions of each feature parameters and the corresponding logical rules are defined by constructing probability distribution histogram and the one dimensional range distribution of the corresponding feature parameters. Several groups of typical Level II data are collected to test and verify the proposed method, the experimental results demonstrate the reliability of the proposed algorithm.
2016, 38(12): 3261-3268.
doi: 10.11999/JEIT160722
Abstract:
Making full and effective use of target polarization information from High Resolution Range Profile (HRRP) is a hot issue for improving the recognition performance of maritime surveillance radar. A HRRP database with seven maritime targets classes from various aspect angles is established, on which thirty-nine features from four categories are defined. A novel feature selection method based on the Normalized Mutual Information (NMI) and Simulated Annealing (SA) algorithm is presented, named as NMI-SA. The effectiveness of the NMI-SA is proved by comparison with three other methods using HRRP dataset and eight from UCI machine learning repository. Finally, the NMI-SA is applied to the HRRP dataset to find twenty-five high discriminant and low redundancy features.
Making full and effective use of target polarization information from High Resolution Range Profile (HRRP) is a hot issue for improving the recognition performance of maritime surveillance radar. A HRRP database with seven maritime targets classes from various aspect angles is established, on which thirty-nine features from four categories are defined. A novel feature selection method based on the Normalized Mutual Information (NMI) and Simulated Annealing (SA) algorithm is presented, named as NMI-SA. The effectiveness of the NMI-SA is proved by comparison with three other methods using HRRP dataset and eight from UCI machine learning repository. Finally, the NMI-SA is applied to the HRRP dataset to find twenty-five high discriminant and low redundancy features.
2016, 38(12): 3269-3274.
doi: 10.11999/JEIT160672
Abstract:
Space-Based Radar (SBR) can realize moving target detection and the high resolution imaging on a global scale. Based on large scale sparse array antenna, this paper discusses the space-based radar technology system and key technology. The system parameters are designed and its performance is analyzed for an X band MEO space-based radar. The large scale sparse array antenna provides a new selection for space-based radar.
Space-Based Radar (SBR) can realize moving target detection and the high resolution imaging on a global scale. Based on large scale sparse array antenna, this paper discusses the space-based radar technology system and key technology. The system parameters are designed and its performance is analyzed for an X band MEO space-based radar. The large scale sparse array antenna provides a new selection for space-based radar.
2016, 38(12): 3275-3281.
doi: 10.11999/JEIT160919
Abstract:
The SMeared SPectrum (SMSP) jamming is a special active deception jamming for countering a Linear Frequency Modulated (LFM) pulse compression radar system. For suppressing the SMSP jamming, a novel anti-jamming approach is proposed based on jointing Time-Frequency (TF) distribution and Compressed Sensing (CS). First the SMSP jamming is isolated in the TF domain, and then the TF points belonging to target signal are reserved. According to the linear relationship between the reserved TF points and the frequency domain of radar echo, a CS minimization model is established and the Orthogonal Matching Pursuit (OMP) algorithm is utilized to recover the target signal, which means the SMSP jamming is mitigated simultaneously. The validity of the proposed method is evaluated using experimental data via Monte Carlo simulation.
The SMeared SPectrum (SMSP) jamming is a special active deception jamming for countering a Linear Frequency Modulated (LFM) pulse compression radar system. For suppressing the SMSP jamming, a novel anti-jamming approach is proposed based on jointing Time-Frequency (TF) distribution and Compressed Sensing (CS). First the SMSP jamming is isolated in the TF domain, and then the TF points belonging to target signal are reserved. According to the linear relationship between the reserved TF points and the frequency domain of radar echo, a CS minimization model is established and the Orthogonal Matching Pursuit (OMP) algorithm is utilized to recover the target signal, which means the SMSP jamming is mitigated simultaneously. The validity of the proposed method is evaluated using experimental data via Monte Carlo simulation.
2016, 38(12): 3282-3288.
doi: 10.11999/JEIT160141
Abstract:
In order to overcome the influence of gray difference, rotation difference and scale difference on image registration accuracy, the gray statistic property of uniform regions in SAR images is utilized and a SAR image registration algorithm based on stable closed uniform regions is proposed. Firstly, based on the multi-scale nonlinear diffusion theory, closed uniform regions with good contour pervserving ability are respectively extracted from two images. Secondly, two affine-invariant region features based on polygon fitting and coincidence degree are constructed to realized the coarse-to-fine region matching. Finally, the centroids of matched regions are used to construct the transform model between two images. Experimental results demonstrate that, the proposed algorithm has high registration accuracy, and is effective for gray difference, rotation difference and scale difference, moreover, it has high adaptability to noise.
In order to overcome the influence of gray difference, rotation difference and scale difference on image registration accuracy, the gray statistic property of uniform regions in SAR images is utilized and a SAR image registration algorithm based on stable closed uniform regions is proposed. Firstly, based on the multi-scale nonlinear diffusion theory, closed uniform regions with good contour pervserving ability are respectively extracted from two images. Secondly, two affine-invariant region features based on polygon fitting and coincidence degree are constructed to realized the coarse-to-fine region matching. Finally, the centroids of matched regions are used to construct the transform model between two images. Experimental results demonstrate that, the proposed algorithm has high registration accuracy, and is effective for gray difference, rotation difference and scale difference, moreover, it has high adaptability to noise.
2016, 38(12): 3289-3297.
doi: 10.11999/JEIT160992
Abstract:
Polarimetric Synthetic Aperture Radar (Polarimetric SAR) has become a hot research topic in the field of remote sensing with the rapid development in recent years. Polarimetric target decomposition is a basic method for Polarimetric SAR image analysis, and plays a key role in Polarimetric SAR image interpretation, the extracted features from polarimetric target decomposition is the basis of target detection and image classification using Polarimetric SAR image. In this paper, through expositing the development of polarimetric target decomposition as well as the new technologies in recent years comprehensively, the relevant researchers can understand the latest progress in this field clearly.
Polarimetric Synthetic Aperture Radar (Polarimetric SAR) has become a hot research topic in the field of remote sensing with the rapid development in recent years. Polarimetric target decomposition is a basic method for Polarimetric SAR image analysis, and plays a key role in Polarimetric SAR image interpretation, the extracted features from polarimetric target decomposition is the basis of target detection and image classification using Polarimetric SAR image. In this paper, through expositing the development of polarimetric target decomposition as well as the new technologies in recent years comprehensively, the relevant researchers can understand the latest progress in this field clearly.
2016, 38(12): 3298-3306.
doi: 10.11999/JEIT161007
Abstract:
Airborne early warning radar and its signal processing technology have experienced great development, but it is also facing great challenges on stealth target, heterogeneous clutter, complex electromagnetic environment, target classification and a variety of combat missions. In this paper, the development of airborne early warning radar and its signal processing technology is reviewed and the challenges airborne early warning radar facing such as anti-stealth, anti-jamming, anti-clutter and target recognition are analyzed. The development trends of the airborne early warning radar system towards digital, broadband, collaborative and intelligent direction are put forward on this basis. Finally, the key technologies of the signal processing such as 3D-STAP, MIMO-STAP, wideband detection, cognitive anti-jamming are analyzed, which have certain directive significance for the development of the next generation airborne early warning radar.
Airborne early warning radar and its signal processing technology have experienced great development, but it is also facing great challenges on stealth target, heterogeneous clutter, complex electromagnetic environment, target classification and a variety of combat missions. In this paper, the development of airborne early warning radar and its signal processing technology is reviewed and the challenges airborne early warning radar facing such as anti-stealth, anti-jamming, anti-clutter and target recognition are analyzed. The development trends of the airborne early warning radar system towards digital, broadband, collaborative and intelligent direction are put forward on this basis. Finally, the key technologies of the signal processing such as 3D-STAP, MIMO-STAP, wideband detection, cognitive anti-jamming are analyzed, which have certain directive significance for the development of the next generation airborne early warning radar.
