Most Downloaded

1
Spaceborne SAR, which is a kind of initiatively microwave imaging sensor, plays an important role in gathering information with its capability of all-day and all-weather imaging, and has become an indispensable sensor for observing the earth. With the development of SAR techniques, Spaceborne SAR has been provided with the ability of High-Resolution Wide-Swath, miniaturization with low cost, bistatic and multi-mode imaging, and Ground Moving Target Indicating (GMTI), so more accurate information about the culture could be obtained with lower cost. In the meantime, more technique problems with muliti-mode, new work system and complex environment are arising and needed to be solved. The main work of this paper is discussing the current situation and the future development of Spaceborne SAR. Spaceborne SAR, which is a kind of initiatively microwave imaging sensor, plays an important role in gathering information with its capability of all-day and all-weather imaging, and has become an indispensable sensor for observing the earth. With the development of SAR techniques, Spaceborne SAR has been provided with the ability of High-Resolution Wide-Swath, miniaturization with low cost, bistatic and multi-mode imaging, and Ground Moving Target Indicating (GMTI), so more accurate information about the culture could be obtained with lower cost. In the meantime, more technique problems with muliti-mode, new work system and complex environment are arising and needed to be solved. The main work of this paper is discussing the current situation and the future development of Spaceborne SAR.
2
The sparse microwave imaging combines the sparse signal processing theory with radar imaging to obtain new theory, new system, and new methodology of microwave imaging. In this paper, a brief review of fundamental issues in applying sparse signal processing to radar imaging is provided, including sparse representation, measurement matrix construction, unambiguity reconstruction, and so on. The developments of sparse signal processing in microwave imaging are discussed, and the initial airborne experiments on the prototype Synthetic Aperture Radar (SAR) framework with sparse constraints are introduced. The results demonstrate the feasibility and effectiveness of the principle and methodology of sparse microwave imaging. Besides, we also provide an overview of sparse signal processing in various radar applications, including Tomographic SAR (TomoSAR), Inverse SAR (ISAR), Ground Penetrating Radar (GPR) as well. The sparse microwave imaging combines the sparse signal processing theory with radar imaging to obtain new theory, new system, and new methodology of microwave imaging. In this paper, a brief review of fundamental issues in applying sparse signal processing to radar imaging is provided, including sparse representation, measurement matrix construction, unambiguity reconstruction, and so on. The developments of sparse signal processing in microwave imaging are discussed, and the initial airborne experiments on the prototype Synthetic Aperture Radar (SAR) framework with sparse constraints are introduced. The results demonstrate the feasibility and effectiveness of the principle and methodology of sparse microwave imaging. Besides, we also provide an overview of sparse signal processing in various radar applications, including Tomographic SAR (TomoSAR), Inverse SAR (ISAR), Ground Penetrating Radar (GPR) as well.
3
In recent years, Internal Waves (IWs) detection in Synthetic Aperture Radar (SAR) image has received considerable attentions in the area of marine remote sensing and has already become one of the most important marine applications of SAR. Typical research results at home and abroad are reviewed. Three areas of researches are introduced and summarized, including parameter inversion method of IWs, the effect of different SAR parameter and wind field conditions on IWs imaging, the 2-dimentional SAR imaging simulation of IWs. In recent years, Internal Waves (IWs) detection in Synthetic Aperture Radar (SAR) image has received considerable attentions in the area of marine remote sensing and has already become one of the most important marine applications of SAR. Typical research results at home and abroad are reviewed. Three areas of researches are introduced and summarized, including parameter inversion method of IWs, the effect of different SAR parameter and wind field conditions on IWs imaging, the 2-dimentional SAR imaging simulation of IWs.
4
Starting from the detection principle and characteristics of passive radar, this paper describes the development of passive radar based on the low frequency band (HF/VHF/UHF) digital broadcasting and TV signal. Based on the radio coverage ratio and technical features of digital broadcasting and TV signals, the research status in abroad, especially in Europe, is introduced at first, on experimental systems, technical parameters, and comparative experiments. Then the latest development of passive radars, in different frequency bands in China, both theory and experimental study are presented. Followed is the commentary on the key techniques and problems of Digital Broadcasting-based Passive Radar (DBPR), including the waveforms properties and its modification, reference signal extraction, multipath clutter rejection, target detection, tracking, and fusion as well as real-time signal processing. Finally, the prospects of development and application of this kind of passive radar are discussed. Starting from the detection principle and characteristics of passive radar, this paper describes the development of passive radar based on the low frequency band (HF/VHF/UHF) digital broadcasting and TV signal. Based on the radio coverage ratio and technical features of digital broadcasting and TV signals, the research status in abroad, especially in Europe, is introduced at first, on experimental systems, technical parameters, and comparative experiments. Then the latest development of passive radars, in different frequency bands in China, both theory and experimental study are presented. Followed is the commentary on the key techniques and problems of Digital Broadcasting-based Passive Radar (DBPR), including the waveforms properties and its modification, reference signal extraction, multipath clutter rejection, target detection, tracking, and fusion as well as real-time signal processing. Finally, the prospects of development and application of this kind of passive radar are discussed.
5
In modern high resolution SAR data, due to the intrinsic side-looking geometry of SAR sensors, layover and foreshortening issues inevitably arise, especially in dense urban areas. SAR tomography provides a new way of overcoming these problems by exploiting the back-scattering property for each pixel. However, traditional non-parametric spectral estimators, e.g. Truncated Singular Value Decomposition (TSVD), are limited by their poor elevation resolution, which is not comparable to the azimuth and slant-range resolution. In this paper, the Compressive Sensing (CS) approach using Basis Pursuit (BP) and TWo-step Iterative Shrinkage/Thresholding (TWIST) are introduced. Experimental studies with real spotlight-mode TerraSAR-X dataset are carried out using both BP and TWIST, to demonstrate the merits of compressive sensing approaches in terms of robustness, computational efficiency, and super-resolution capability. In modern high resolution SAR data, due to the intrinsic side-looking geometry of SAR sensors, layover and foreshortening issues inevitably arise, especially in dense urban areas. SAR tomography provides a new way of overcoming these problems by exploiting the back-scattering property for each pixel. However, traditional non-parametric spectral estimators, e.g. Truncated Singular Value Decomposition (TSVD), are limited by their poor elevation resolution, which is not comparable to the azimuth and slant-range resolution. In this paper, the Compressive Sensing (CS) approach using Basis Pursuit (BP) and TWo-step Iterative Shrinkage/Thresholding (TWIST) are introduced. Experimental studies with real spotlight-mode TerraSAR-X dataset are carried out using both BP and TWIST, to demonstrate the merits of compressive sensing approaches in terms of robustness, computational efficiency, and super-resolution capability.
6
Global Navigation Satellite System (GNSS), has a significant impact on all areas of human activity, not only can provide users with shared global navigation, position and timing information, but also can provide a L-band microwave signal source of long term stability and high temporal-spatial resolution. In recent years, development of the navigation satellite remote sensing applications using GNSS as a external illuminator, it has been forming a new Global Navigation Satellite System METeorology (GNSS/MET), of which Global Navigation Satellite System-Reflection (GNSS-R) signals remote sensing technology is rising. It could be considered as a non-cooperative artificial illuminator, bistatic (multi-static) radar system, and has the advantages of both passive and active remote sensing. Then it gets more and more peoples attention and favor, and broadening into Atmosphere -ocean and land surface remote sensing fields. However, the address of this technology is very messy at home and abroad, and not able to accurately express its special meaning. This article attempts to give a new term: Exogenous-Aided Remote Sensing (EARS) for discussion. Global Navigation Satellite System (GNSS), has a significant impact on all areas of human activity, not only can provide users with shared global navigation, position and timing information, but also can provide a L-band microwave signal source of long term stability and high temporal-spatial resolution. In recent years, development of the navigation satellite remote sensing applications using GNSS as a external illuminator, it has been forming a new Global Navigation Satellite System METeorology (GNSS/MET), of which Global Navigation Satellite System-Reflection (GNSS-R) signals remote sensing technology is rising. It could be considered as a non-cooperative artificial illuminator, bistatic (multi-static) radar system, and has the advantages of both passive and active remote sensing. Then it gets more and more peoples attention and favor, and broadening into Atmosphere -ocean and land surface remote sensing fields. However, the address of this technology is very messy at home and abroad, and not able to accurately express its special meaning. This article attempts to give a new term: Exogenous-Aided Remote Sensing (EARS) for discussion.
7
This paper first reviews the history and trends in the development of spaceborne Synthetic Aperture Radar (SAR) satellite technology in the USA and Europe. The basic information regarding launched satellites and future satellite plans are introduced. Then, this paper summarizes and categorizes the imaging algorithms of spaceborn SAR satellites, and analyzes the advantages and disadvantages of each algorithm. Next, the scope and the application status of each algorithm are presented. Then, the paper presents details of trends related to the SAR imaging algorithm, which mainly introduces the algorithms based on compressive sensing theory and new image modes. The simulation results are also presented. Finally, we summarize the development direction of the spaceborne SAR imaging algorithm. This paper first reviews the history and trends in the development of spaceborne Synthetic Aperture Radar (SAR) satellite technology in the USA and Europe. The basic information regarding launched satellites and future satellite plans are introduced. Then, this paper summarizes and categorizes the imaging algorithms of spaceborn SAR satellites, and analyzes the advantages and disadvantages of each algorithm. Next, the scope and the application status of each algorithm are presented. Then, the paper presents details of trends related to the SAR imaging algorithm, which mainly introduces the algorithms based on compressive sensing theory and new image modes. The simulation results are also presented. Finally, we summarize the development direction of the spaceborne SAR imaging algorithm.
8
Bistatic SAR (BiSAR) systems have attracted the interests from global researchers and become a hotspot in the international radar community due to the progress of radar technology and rapidly increased applications nowadays. Based on the BiSAR experiments and breakthrough of the key technology, the paper summarized the general progresses of BiSAR systems, especially in European radar community, from different aspects such as system design, processing idea and topology etc. Different bistatic image formation algorithms have been analyzed and reviewed. Finally, the development trend is discussed in the paper. Bistatic SAR (BiSAR) systems have attracted the interests from global researchers and become a hotspot in the international radar community due to the progress of radar technology and rapidly increased applications nowadays. Based on the BiSAR experiments and breakthrough of the key technology, the paper summarized the general progresses of BiSAR systems, especially in European radar community, from different aspects such as system design, processing idea and topology etc. Different bistatic image formation algorithms have been analyzed and reviewed. Finally, the development trend is discussed in the paper.
9
Viewing from the interaction between external and internal causes on the time scale of history, present and future, this paper analyzes and demonstrates the developing motivation and stage characteristics of radar technology. The external causes are interpreted as target, environment and mission, and the internal causes as information acquisition pattern, realization ability and resource utilization. The fundamental law of radar development is revealed as evolving stepwise from lower into higher dimension of detection through the aromorphosis of channel configuration, viewing angle and signal dimensionality, while the main innovation strategies of radar technology are summarized as modifying information acquisition pattern, enhancing realization ability and increasing utilized resources. Furthermore, the developing trends and main characteristics of future radar technology are deduced, and proposals for promoting future innovation and development are also presented. Viewing from the interaction between external and internal causes on the time scale of history, present and future, this paper analyzes and demonstrates the developing motivation and stage characteristics of radar technology. The external causes are interpreted as target, environment and mission, and the internal causes as information acquisition pattern, realization ability and resource utilization. The fundamental law of radar development is revealed as evolving stepwise from lower into higher dimension of detection through the aromorphosis of channel configuration, viewing angle and signal dimensionality, while the main innovation strategies of radar technology are summarized as modifying information acquisition pattern, enhancing realization ability and increasing utilized resources. Furthermore, the developing trends and main characteristics of future radar technology are deduced, and proposals for promoting future innovation and development are also presented.
10
With the development of high speed digital processor and solid state power electronics, more flexible waveforms become feasible to achieve by modern radar systems. In fact, the choice of waveforms has a significant impact on the performance of radar systems. In this paper, we review the conventional radar waveform design as well as explore the new generation of waveforms via different theoretical methods, including the most recent wavelet based waveforms. It is shown that the waveform design supports the radar advancement for more intelligent and divergent applications. In this endeavor, radar waveform design plays an even more important role in reaching specific purposes, in addition to range and speed detections, and further improves the performance and application scopes of radar systems. With the development of high speed digital processor and solid state power electronics, more flexible waveforms become feasible to achieve by modern radar systems. In fact, the choice of waveforms has a significant impact on the performance of radar systems. In this paper, we review the conventional radar waveform design as well as explore the new generation of waveforms via different theoretical methods, including the most recent wavelet based waveforms. It is shown that the waveform design supports the radar advancement for more intelligent and divergent applications. In this endeavor, radar waveform design plays an even more important role in reaching specific purposes, in addition to range and speed detections, and further improves the performance and application scopes of radar systems.
11
Collision avoidance radar for trains is pregnant for safety transportation. In order to realize low cost and high performance of azimuth accuracy, we have developed MMW (Milli-Meter Wave) radar, which employs switched phased array and frequency stepped technology. This paper analyses the radiation patterns of transmitting/receiving antennas and compensation method for amplitude/phase errors of synthetic wideband frequency stepped signal. To confirm the operation of the radar, low cost millimeter-wave collision avoidance radar was fabricated. Lots of experiments confirmed a high azimuth and range resolution. Collision avoidance radar for trains is pregnant for safety transportation. In order to realize low cost and high performance of azimuth accuracy, we have developed MMW (Milli-Meter Wave) radar, which employs switched phased array and frequency stepped technology. This paper analyses the radiation patterns of transmitting/receiving antennas and compensation method for amplitude/phase errors of synthetic wideband frequency stepped signal. To confirm the operation of the radar, low cost millimeter-wave collision avoidance radar was fabricated. Lots of experiments confirmed a high azimuth and range resolution.
12
The HJ-1C satellite was successfully launched in November 19, 2012. The HJ-1C and HJ-1A/1B satellites, which were launched in September 06, 2008, constitute the 2+1 small satellite constellation for environmental and disaster monitoring. This study focuses on the analysis and evaluation of the satellite performance with respect to environmental remote sensing, including land use interpretation, land cover classification, oil spill identification, retrieval of sea waves, and monitoring of coastal mariculture. The data used in this study cover the city of Beijing and the sea of the Fujian Province. Nine HJ-1C satellite images (level-2, S band, VV Pol, strip mode, 5 m resolution) from December 2012 to January 2013 are used. The conclusions are as follows: (1) the HJ-1C SAR images can be used to manually identify farmland, woodland, roads, rivers, urban construction, and rural residential areas; (2) the accuracy of the automatic land cover classification increased significantly when the HJ-1C SAR and HJ-1B CCD fusion images are used; (3) the HJ-1C satellite can be used to identify oil spills, to invert wave parameters, and to extract information regarding inshore aquaculture. The HJ-1C satellite was successfully launched in November 19, 2012. The HJ-1C and HJ-1A/1B satellites, which were launched in September 06, 2008, constitute the 2+1 small satellite constellation for environmental and disaster monitoring. This study focuses on the analysis and evaluation of the satellite performance with respect to environmental remote sensing, including land use interpretation, land cover classification, oil spill identification, retrieval of sea waves, and monitoring of coastal mariculture. The data used in this study cover the city of Beijing and the sea of the Fujian Province. Nine HJ-1C satellite images (level-2, S band, VV Pol, strip mode, 5 m resolution) from December 2012 to January 2013 are used. The conclusions are as follows: (1) the HJ-1C SAR images can be used to manually identify farmland, woodland, roads, rivers, urban construction, and rural residential areas; (2) the accuracy of the automatic land cover classification increased significantly when the HJ-1C SAR and HJ-1B CCD fusion images are used; (3) the HJ-1C satellite can be used to identify oil spills, to invert wave parameters, and to extract information regarding inshore aquaculture.
13
This paper gives the experimental research of HF Passive Bistatic Radar (HFPBR) based on Digital Radio Mondiale (DRM) digital AM broadcasting that have been first carried out in China, using the newly-developed all-digital active/passive integrated HF surface wave radar system. The principle, key techniques, experimental equipment, and preliminary results are introduced about this new radar system. Based on analysis of the measurement data, experimental results under different scenarios including surface-wave, sky-wave, and hybrid sky-surface propagation modes are presented, which have proved, for the first time worldwide, the technical feasibility of using DRM broadcasting signal for over-the-horizon detection by field experiment and formed the theoretical and experimental basis for the further development of HFPBR. This paper gives the experimental research of HF Passive Bistatic Radar (HFPBR) based on Digital Radio Mondiale (DRM) digital AM broadcasting that have been first carried out in China, using the newly-developed all-digital active/passive integrated HF surface wave radar system. The principle, key techniques, experimental equipment, and preliminary results are introduced about this new radar system. Based on analysis of the measurement data, experimental results under different scenarios including surface-wave, sky-wave, and hybrid sky-surface propagation modes are presented, which have proved, for the first time worldwide, the technical feasibility of using DRM broadcasting signal for over-the-horizon detection by field experiment and formed the theoretical and experimental basis for the further development of HFPBR.
14
The objective of this paper is to investigate the connotation, necessity and tendency of Synthetic Aperture Radar (SAR) imaging technology within the framework of multidimensional space joint-observation, which are polarimetry space, frequency space, angle space and time space, on the basis of the key evolvement phases of SAR imaging technology. Furthermore,the definition of Multidimensional Space Joint-observation SAR (MSJosSAR) is given based on the demand of information fusion of multidimensional space joint observation SAR images. After that, the advantage of MSJosSAR is revealed by using Kronecker product decomposition technology for better understanding of target scattering mechanisms. Besides, the hypothesis and basic framework on which the MSJosSAR signal processing rely is listed and illustrated. Finally, the number of joint observation spaces of typical SAR configurations is enumerated. The objective of this paper is to investigate the connotation, necessity and tendency of Synthetic Aperture Radar (SAR) imaging technology within the framework of multidimensional space joint-observation, which are polarimetry space, frequency space, angle space and time space, on the basis of the key evolvement phases of SAR imaging technology. Furthermore,the definition of Multidimensional Space Joint-observation SAR (MSJosSAR) is given based on the demand of information fusion of multidimensional space joint observation SAR images. After that, the advantage of MSJosSAR is revealed by using Kronecker product decomposition technology for better understanding of target scattering mechanisms. Besides, the hypothesis and basic framework on which the MSJosSAR signal processing rely is listed and illustrated. Finally, the number of joint observation spaces of typical SAR configurations is enumerated.
15
As a new radar technology, the distributed aperture coherent radar is expected to be the next generation radar, which is easier to transport and less expensive than the traditional large aperture radar. However, the time synchronization and phase synchronization are key issues to be addressed for the distributed aperture coherent radar. In this paper, the error sources of time synchronization and phase synchronization are analyzed, and the corresponding mathematical models are first derived. Then, the impact of synchronization errors on the coherent performance is simulated, and the accuracy of time and phase synchronization is presented based on the simulation results. Finally, the noncorrelation transmission scheme and the calibration scheme based on the wired transmission are proposed to realize the time and phase synchronization, respectively. Research of the synchronization problem could be very helpful to realize the new radar technology of distributed aperture coherent radar. As a new radar technology, the distributed aperture coherent radar is expected to be the next generation radar, which is easier to transport and less expensive than the traditional large aperture radar. However, the time synchronization and phase synchronization are key issues to be addressed for the distributed aperture coherent radar. In this paper, the error sources of time synchronization and phase synchronization are analyzed, and the corresponding mathematical models are first derived. Then, the impact of synchronization errors on the coherent performance is simulated, and the accuracy of time and phase synchronization is presented based on the simulation results. Finally, the noncorrelation transmission scheme and the calibration scheme based on the wired transmission are proposed to realize the time and phase synchronization, respectively. Research of the synchronization problem could be very helpful to realize the new radar technology of distributed aperture coherent radar.
16
In this paper, the definition and the key features of Software Radar, which is a new concept, are proposed and discussed. We consider the development of modern radar system technology to be divided into three stages: Digital Radar, Software radar and Intelligent Radar, and the second stage is just commencing now. A Software Radar system should be a combination of various modern digital modular components conformed to certain software and hardware standards. Moreover, a software radar system with an open system architecture supporting to decouple application software and low level hardware would be easy to adopt user requirements-oriented developing methodology instead of traditional specific function-oriented developing methodology. Compared with traditional Digital Radar, Software Radar system can be easily reconfigured and scaled up or down to adapt to the changes of requirements and technologies. A demonstration Software Radar signal processing system, RadarLab 2.0, which has been developed by Tsinghua University, is introduced in this paper and the suggestions for the future development of Software Radar in China are also given in the conclusion. In this paper, the definition and the key features of Software Radar, which is a new concept, are proposed and discussed. We consider the development of modern radar system technology to be divided into three stages: Digital Radar, Software radar and Intelligent Radar, and the second stage is just commencing now. A Software Radar system should be a combination of various modern digital modular components conformed to certain software and hardware standards. Moreover, a software radar system with an open system architecture supporting to decouple application software and low level hardware would be easy to adopt user requirements-oriented developing methodology instead of traditional specific function-oriented developing methodology. Compared with traditional Digital Radar, Software Radar system can be easily reconfigured and scaled up or down to adapt to the changes of requirements and technologies. A demonstration Software Radar signal processing system, RadarLab 2.0, which has been developed by Tsinghua University, is introduced in this paper and the suggestions for the future development of Software Radar in China are also given in the conclusion.
17
The concept of Geosynchronous Circular SAR (Geo-CSAR) is introduced in this paper. With the design of the geosynchronous orbit parameters, a near-circular satellite sub-track could be formed to enable the staring imaging mode, which supports the advanced applications for wide-field and 3-D information acquisition under long-term consistent observation. This paper also analyzes Geo-CSAR's imaging formation capabilities, and concludes its attractive advantages over low-earth orbit spaceborne SAR in terms of instantaneous coverage, consistent observing area, 3-D positioning accuracy and etc.. Encouraging expectations for Geo-CSAR thus could be positively predicted in military investigation and disaster monitoring management applications. The concept of Geosynchronous Circular SAR (Geo-CSAR) is introduced in this paper. With the design of the geosynchronous orbit parameters, a near-circular satellite sub-track could be formed to enable the staring imaging mode, which supports the advanced applications for wide-field and 3-D information acquisition under long-term consistent observation. This paper also analyzes Geo-CSAR's imaging formation capabilities, and concludes its attractive advantages over low-earth orbit spaceborne SAR in terms of instantaneous coverage, consistent observing area, 3-D positioning accuracy and etc.. Encouraging expectations for Geo-CSAR thus could be positively predicted in military investigation and disaster monitoring management applications.
18
Circular SAR (CSAR) is a newly developed all-directional high resolution 3D imaging mode in recent years, to satisfy the demand of finer observation. The National Key Laboratory of Science and Technology on Microwave Imaging, Institute of Electronics, Chinese Academy of Sciences (MITL, IECAS), had the first test flight experiment in Aug. 2011 with a P-band full polarization SAR system, and successfully obtained the all-directional high resolution circular SAR image. The initial results show that CSAR technique has the encouraging potential capability in the fields of high precision mapping, disaster evaluation, resource management and the other related applications. This paper firstly makes a detailed discussion on the progress of circular SAR imaging technique, which emphases on the several airborne experiments performed these years to show CSARs attractive features, then studies and illustrates the key techniques, and finally discusses the development trends. Circular SAR (CSAR) is a newly developed all-directional high resolution 3D imaging mode in recent years, to satisfy the demand of finer observation. The National Key Laboratory of Science and Technology on Microwave Imaging, Institute of Electronics, Chinese Academy of Sciences (MITL, IECAS), had the first test flight experiment in Aug. 2011 with a P-band full polarization SAR system, and successfully obtained the all-directional high resolution circular SAR image. The initial results show that CSAR technique has the encouraging potential capability in the fields of high precision mapping, disaster evaluation, resource management and the other related applications. This paper firstly makes a detailed discussion on the progress of circular SAR imaging technique, which emphases on the several airborne experiments performed these years to show CSARs attractive features, then studies and illustrates the key techniques, and finally discusses the development trends.
19
Automatic Target Recognition (ATR) of Synthetic Aperture Radar (SAR) image is investigated. A SAR feature extraction algorithm based on multilayer auto-encoder is proposed. The method makes use of a probabilistic neural network, Restricted Boltzmann Machine (RBM) modeling probability distribution of environment. Through the formation of more expressive multilayer neural network, the deep learning model learns shared representation of the target and its shadow outline reflecting the target shape characteristics. Targets are classified automatically through two recognition models. The experiment results based on the MSTAR verify the effectiveness of proposed algorithm. Automatic Target Recognition (ATR) of Synthetic Aperture Radar (SAR) image is investigated. A SAR feature extraction algorithm based on multilayer auto-encoder is proposed. The method makes use of a probabilistic neural network, Restricted Boltzmann Machine (RBM) modeling probability distribution of environment. Through the formation of more expressive multilayer neural network, the deep learning model learns shared representation of the target and its shadow outline reflecting the target shape characteristics. Targets are classified automatically through two recognition models. The experiment results based on the MSTAR verify the effectiveness of proposed algorithm.
20
To improve the computational efficiency of system feature extraction, reduce the occupied memory space, and simplify the program design, a modified gradient descent method on Stiefel manifold is proposed based on the optimization algorithm of geometry frame on the Riemann manifold. Different geodesic calculation formulas are used for different scenarios. A polynomial is also used to lie close to the geodesic equations. JiuZhaoQin-Horner polynomial algorithm and the strategies of line-searching technique and change of the step size of iteration are also adopted. The gradient descent algorithm on Stiefel manifold applied in Principal Component Analysis (PCA) is discussed in detail as an example of system feature extraction. Theoretical analysis and simulation experiments show that the new method can achieve superior performance in both the convergence rate and calculation efficiency while ensuring the unitary column orthogonality. In addition, it is easier to implement by software or hardware. To improve the computational efficiency of system feature extraction, reduce the occupied memory space, and simplify the program design, a modified gradient descent method on Stiefel manifold is proposed based on the optimization algorithm of geometry frame on the Riemann manifold. Different geodesic calculation formulas are used for different scenarios. A polynomial is also used to lie close to the geodesic equations. JiuZhaoQin-Horner polynomial algorithm and the strategies of line-searching technique and change of the step size of iteration are also adopted. The gradient descent algorithm on Stiefel manifold applied in Principal Component Analysis (PCA) is discussed in detail as an example of system feature extraction. Theoretical analysis and simulation experiments show that the new method can achieve superior performance in both the convergence rate and calculation efficiency while ensuring the unitary column orthogonality. In addition, it is easier to implement by software or hardware.
  • First
  • Prev
  • 1
  • 2
  • 3
  • 4
  • 5
  • Last
  • Total:5
  • To
  • Go