Recently, frequency-modulated continuous-wave (FMCW) radar-based hand gesture recognition (HGR) using deep learning has achieved favorable performance. However, many existing methods use extracted features separately, i.e., using one of the range, Doppler, azimuth, or elevation angle information, or a combination of any two, to train convolutional neural networks (CNNs), which ignore the interrelation among the 5-D time-varying-range-Doppler-azimuth-elevation feature space. Although there have been methods using the 5-D information, their mining of the interrelation among the 5-D feature space is not sufficient, and there is still room for improvements. This article proposes a new processing scheme of HGR based on 5-D feature cubes that are jointly encoded by a 3-D fast Fourier transform (3-D-FFT)-based method. Then, a CNN is proposed by building two novel blocks, i.e., the spatiotemporal deformable convolution (STDC) block and the adaptive spatiotemporal context-aware convolution (ASTCAC) block. Concretely, STDC is designed to cope with hand gestures' large spatiotemporal geometric transformations in the 5-D feature space. Moreover, ASTCAC is designed for modeling long-distance global relationships, e.g., relationships between pixels of the feature at the upper left corner and lower right corner, and exploring the global spatiotemporal context, in order to enhance the target feature representation and suppress interference. Finally, our presented method is verified on a large radar dataset, including 19 760 sets of 16 common hand gestures, collected by 19 subjects. Our method obtains a recognition rate of 99.53% on the validation dataset and that of 97.22% on the test dataset, which is significantly better than state-of-the-art methods.

@en

Nowadays deep learning-based weather radar echo extrapolation methods have competently improved nowcasting quality. Current pure convolutional or convolutional recurrent neural network-based extrapolation pipelines inherently struggle in capturing both global and local spatiotemporal interactions simultaneously, thereby limiting nowcasting performances, e.g., they not only tend to underestimate heavy rainfalls' spatial coverage and intensity but also fail to precisely predict nonlinear motion patterns. Furthermore, the usually adopted pixel-wise objective functions lead to blurry predictions. To this end, we propose a novel motion-guided global-local aggregation Transformer network for effectively combining spatiotemporal cues at different time scales, thereby strengthening global-local spatiotemporal aggregation urgently required by the extrapolation task. First, we divide existing observations into both short- and long-term sequences to represent echo dynamics at different time scales. Then, to introduce reasonable motion guidance to Transformer, we customize an end-to-end module for jointly extracting motion representation of short- and long-term echo sequences (MRS, MRL), while estimating optical flow. Subsequently, based on Transformer architecture, MRS is used as queries to retrospect the most useful information from MRL for an effective aggregation of global long-term and local short-term cues. Finally, the fused feature is employed for future echo prediction. Additionally, for the blurry prediction problem, predictions from our model trained with an adversarial regularization achieve superior performances not only in nowcasting skill scores but also in precipitation details and image clarity over existing methods. Extensive experiments on two challenging radar echo datasets demonstrate the effectiveness of our proposed method.

@en

Coherence-based synthetic aperture radar (SAR) tomography (TomoSAR) exploits the complex coherences of SAR images to achieve 3-D imaging. Utilizing two-sensor spaceborne SAR formation flying to realize coherence-based TomoSAR has attracted increasing attention because temporal decorrelation-free interferograms can be constructed; therefore, TomoSAR has excellent potential for inverting the vertical structures of natural scenes such as forests and glaciers. However, low earth orbit (LEO) TomoSAR is disadvantaged by limited data and nonuniform sampling in the elevation direction. Geosynchronous (GEO) TomoSAR can overcome these limitations owing to its short revisit time of no more than 24 h. For the first time, this article discusses coherence-based TomoSAR exploiting GEO SAR formation flying. The benefits of GEO-formation coherence-based TomoSAR, including the low cost of slave satellites, rich data sets, and uniform sampling, are noted. The key problems of system design, including the formation design and data acquisition, are discussed. A formation design method based on the minimum along-track baseline is proposed that can realize uniform elevation sampling. The geometric correlation of a general SAR observation geometry is derived; on this basis, an optimal data acquisition method based on the optimal height measurement Cramer-Rao lower bound (CRLB) is proposed. Finally, the performance of GEO-formation coherence-based TomoSAR is analyzed; in particular, the ambiguity height in the altitude direction, the Rayleigh resolution in the altitude direction, and the theoretical optimal geometric correlation are evaluated. Finally, computer simulations validate the proposed formation design method, data acquisition scheme, and performance analysis formula.

@en

Based on its ability to obtain two-dimensional (2D) high-resolution images in all-time and all-weather conditions, spaceborne synthetic aperture radar (SAR) has become an important remote sensing technique and the study of such systems has entered a period of vigorous development. Advanced imaging modes such as radar interferometry, tomography, and multi-static imaging, have been demonstrated. However, current in-orbit spaceborne SARs, which all operate in low Earth orbits, have relatively long revisit times ranging from several days to dozens of days, restricting their temporal sampling rate. Geosynchronous SAR (GEO SAR) is an active research area because it provides significant new capability, especially its much-improved temporal sampling. This paper reviews the research progress of GEO SAR technologies in detail. Two typical orbit schemes are presented, followed by the corresponding key issues, including system design, echo focusing, main disturbance factors, repeat-track interferometry, etc, inherent to these schemes. Both analysis and solution research of the above key issues are described. GEO SAR concepts involving multiple platforms are described, including the GEO SAR constellation, GEO-LEO/airborne/unmanned aerial vehicle bistatic SAR, and formation flying GEO SAR (FF-GEO SAR). Due to the high potential of FF-GEO SAR for three-dimensional (3D) deformation retrieval and coherence-based SAR tomography (TomoSAR), we have recently carried out some research related to FF-GEO SAR. This research, which is also discussed in this paper, includes developing a formation design method and an improved TomoSAR processing algorithm. It is found that GEO SAR will continue to be an active topic in the aspect of data processing and multi-platform concept in the near future.

@en

With the development trends of multistatic spaceborne synthetic aperture radar (SAR), geosynchronous SAR (GEO SAR) employing several formation-flying small satellites also has great potential for remote sensing. The small satellites can cooperate to acquire multi-channel data for moving target detection and parameter estimation in strong clutters. However, multistatic GEO SAR has large satellite spacing and a curved trajectory, which induce the near-field effects and channels out of alignment, respectively, bringing about challenges for the spatial adaptive processing. These problems produce a high-order term in the multi-channel slant range model, making the traditional model and adaptive processing method invalid. In this paper, to meet the requirement of SAR focusing, we firstly derive a fourth-order slant range model and a third-order path difference model for multistatic GEO SAR. Secondly, based on the derived model, the principle of stationary phase and series reversion method are utilized to derive the spatial steering vector for a moving target, which is a basis of spatial adaptive processing in the range-Doppler domain. Thirdly, the time-domain match filtering is constructed based on the fourth-order slant range model to image the moving target. Additionally, the moving targets are detected in the image domain. The motion parameter is estimated by iteratively maximizing the output signal to clutter and noise ratio (SCNR) through the range of possible target velocities. Finally, considering that the GEO SAR is still in development, the computer simulations are carried out to verify the effectiveness and evaluate the performance.@en

Geosynchronous spaceborne-airborne bistatic synthetic aperture radar (GEO SA-BSAR), consisting of GEO transmitter and airborne receiver, has stable coverage for a long time and benefits moving target detection. However, the performance of GEO SA-BSAR moving target indication (MTI) system varies widely between bistatic configurations. The traditional configuration design for GEO SA-BSAR system only considers the imaging performance, which may cause the poor MTI performance. In this paper, we propose a bistatic configuration design method to jointly optimize the MTI and SAR imaging performance for GEO SA-BSAR MTI system. The relationship between the MTI performance and bistatic configuration parameters is derived analytically and analyzed based on the maximum output signal to clutter and noise ratio (SCNR) criterion. Then, the MTI performance and SAR imaging performance are jointly considered to model the configuration design problem as a multi-objective optimization problem under the constrained condition. Finally, the optimal configuration for GEO SA-BSAR MTI system is given.

@en

The single-pass geosynchronous synthetic aperture radar interferometry (GEO InSAR) adopts the formation of a slave satellite accompanying the master satellite, which can reduce the temporal decorrelation caused by atmospheric disturbance and observation time gap between repeated tracks. Current formation design methods for spaceborne SAR are based on the Relative Motion Equation (RME) in the Earth-Centered-Inertial (ECI) coordinate system (referred to as ECI-RME). Since the Earth rotation is not taken into account, the methods will lead to a significant error for the baseline calculation while applied to formation design for GEO InSAR. In this paper, a formation design method for single-pass GEO InSAR based on Coordinate Rotational Transformation (CRT) is proposed. Through CRT, the RME in Earth-Centered-Earth-Fixed (ECEF) coordinate system (referred to as ECEF-RME) is derived. The ECEF-RME can be used to describe the accurate baseline of close-flying satellites for different orbital altitudes, but not limited to geosynchronous orbit. Aiming at the problem that ECEF-RME does not have a regular geometry as ECI-RME does, a numerical formation design method based on the minimum baseline error criterion is proposed. Then, an analytical formation design method is proposed for GEO InSAR, based on the Minimum Along-track Baseline Criterion (MABC) subject to a fixed root mean square of the perpendicular baseline. Simulation results verify the validity of the ECEF-RME and the analytical formation design method. The simulation results also show that the proposed method can help alleviate the atmospheric phase impacts and improve the retrieval accuracy of the digital elevation model (DEM) compared with the ECI-RME-based approach.

@en

Despite the increasing interest in geosynchronous synthetic aperture radar (GEO SAR) systems, the ground moving target indication and motion parameters estimation aspect have never been addressed in GEO SAR scenarios. In this article, we tackle the issue of multiple moving target velocity estimation in GEO SAR. We develop new closed-form expressions that relate both Doppler centroid and Doppler rate to the target motion parameters in GEO SAR by considering the specific features of a geosynchronous orbit. Furthermore, we propose a new velocity estimation algorithm that combines the nonuniform cubic phase function (NU-CPF) algorithm with the newly developed models to estimate the moving target's two velocity components. Moreover, based on the abovementioned technique, we further propose a solution to address the multiple moving targets' problem. Simulation results along with an estimation accuracy analysis are provided to demonstrate the effectiveness of the proposed multitarget GEO SAR velocity estimation technique.

@en

The sub-satellite track of geosynchronous synthetic aperture radar (GEO SAR) presents the figure "8" or "O", which causes the great changes of platform motion direction and the different projection of anisotropic irregularities along the line-of-sight (LOS) direction. Due to the almost equal angle velocity to that of Earth, the GEO SAR has smaller ionospheric penetration point (IPP) scanning velocity which is much smaller to the counterpart of the low earth orbit SAR (LEO SAR) while is comparable to the drifting velocity of irregularities, which will affect the effective azimuthal velocity. These facts lead to the consequence that the satellite signals from the GEO SAR would become more vulnerable when they are transmitted in the environment where the ionospheric scintillation occurs. However, few works are focused on these mentioned issues towards the GEO SAR system. In this paper, the impacts of ionospheric scintillation on GEO SAR imaging will be analyzed considering the anisotropy and drifting velocity of irregularities. The anisotropy and drifting velocity effects can both originate from the effect on power spectral density (PSD) of phase screen which is used to model the ionospheric scintillation effects. Based on the data from international geomagnetic reference field (IGRF) and satellite tool kit (STK), the GEO SAR imaging simulations for different GEO SAR orbital configurations and positions are carried out. The simulation results demonstrate that the anisotropy and the drifting velocity of irregularities will cause the changes of stripe direction and affect the quality of GEO SAR images.

@en

Geosynchronous orbit synthetic aperture radar (GEO SAR) has a long integration time and a large imaging scene. Therefore, various nonideal factors are easily accumulated, introducing phase errors and degrading the imaging quality. Within the long integration time, tropospheric status changes with time and space, which will result in image shifts and defocusing. According to the characteristics of GEO SAR, the modeling, and quantitative analysis of background troposphere and turbulence are conducted. For background troposphere, the accurate GEO SAR signal spectrum, which takes into account the time-varying troposphere, is deduced. The influences of different rates of changing (ROC) of troposphere with time are analyzed. Finally, results are verified using the refractive index profile data from Fengyun (FY) 3C satellite and the tropospheric zenith delays data from international GNSS service (IGS). The time-space changes of troposphere can cause image shifts which only depend on the satellite beam-foot velocity and the linear ROC of troposphere. The image defocusing is related to the wavelength, resolution requirement, and the second and higher orders of ROC. The short-wavelength GEO SAR systems are more susceptible to impacts, while L-band GEO SAR will be affected when the integration time becomes longer. Tropospheric turbulence will cause the amplitude and phase random fluctuations resulting in image defocusing. However, in the natural environment, radio waves are very weakly affected by turbulence, and the medium-inclined GEO SAR of L- to C-band will not be affected, while the Xband will be influenced slightly.@en