Reliable localization under Global Positioning System-denied or visually degraded conditions remains a fundamental challenge for autonomous systems. Vision- and Light Detection and Ranging (LiDAR)-based approaches often degrade in low illumination, adverse weather, or appearance-changing environments, as they rely on stable surface texture or geometry. In contrast, ground-penetrating radar (GPR) captures subsurface electromagnetic reflections that remain relatively stable across lighting, seasonal, and weather variations, making it a promising complementary sensing modality for long-term localization. However, spatial variability in subsurface dielectric properties induces fluctuations in electromagnetic wave velocity, leading to geometric distortions in GPR echoes and unstable feature extraction. To address this challenge, we propose the Velocity-Invariant Feature Transform (VIFT), a physics-guided self-supervised learning framework for GPR keypoint detection and description. VIFT explicitly models wave-velocity-induced distortions through a continuous velocity space parameterized by a Beta distribution, and leverages velocity-conditioned wavefield migration as physically consistent data augmentation. A Siamese network is trained with velocity-consistency supervision to jointly learn repeatable keypoint score maps and discriminative local descriptors from unlabeled real GPR scans. To further enhance robustness, sparsity-aware, dispersion, distinctiveness, and orthogonality losses are incorporated to improve repeatability, spatial coverage, and descriptor discriminability. Extensive experiments on public benchmarks and large-scale real-world GPR datasets demonstrate that VIFT consistently outperforms traditional handcrafted methods and recent learning-based Vison and GPR methods, achieving a 5–10% improvement in keypoint repeatability over state-of-the-art methods, particularly under extremely sparse keypoint sampling regimes, while also improving matching accuracy and registration robustness under diverse subsurface conditions.

Ultra-wideband (UWB) radar offers strong penetration capability and high spatiotemporal resolution, making it particularly advantageous for contactless human behavior recognition. However, existing approaches mainly depend on single-channel radar data and micro-Doppler spectrograms, which exhibit degraded robustness and accuracy in complex multi-target environments. To address this limitation, we propose a novel multi-target behavior recognition framework that exploits image-domain micro-Doppler features for joint detection and recognition. Specifically, back-projection (BP) is first employed to generate range–azimuth maps. A YOLOv11-based detector is then applied for precise target detection and tracking, effectively reducing missed detections and false alarms commonly encountered in conventional CFAR methods. Individual micro-Doppler spectrograms are generated through coherent spatiotemporal aggregation and short-time Fourier transform (STFT). Furthermore, a lightweight Residual Wavelet Attention Network (RWANet) is developed to enhance feature representation via multi-level wavelet decomposition and attention mechanisms, while preserving a low parameter count and computational efficiency. Experimental results demonstrate that the proposed method achieves 97.24% accuracy in multi-person behavior recognition, validating its effectiveness and practical applicability.

Using radar to contactlessly monitor physiological information plays an important role in advancing the Internet of Health Things (IoHT) industry. In radar-based physiological monitoring tasks, one key issue is how to sense the chest wall micromotions induced by human respiratory and cardiac activities. Unfortunately, these micromotions, which carry physiological signs, are tiny and hard to capture rapidly with high accuracy, making it challenging to balance speed and accuracy—even with the state-of-the-art (SOTA) linear demodulation (LiDe) approach (most accurate but time-consuming). However, in this paper, we explore and develop an improved LiDe method that uses a simple transform and eliminates the need for the singular value decomposition (SVD) used in the original approach, thereby achieving rapid radar sensing of physiological micromotions with SOTA accuracy. This work begins with the surprising discovery that the second component rather than the principal component is valid in LiDe results, and the reasons behind this discovery are also revealed. Consequently, the original LiDe approach should be corrected (cLiDe). Then, the above exploration inspires us to derive a simple linear operator and develop an efficient version of the cLiDe approach, named the sLiDe method, so that the computational speed can be greatly enhanced while maintaining the same high estimation accuracy as cLiDe. Extensive experiments also validate the superiority of our proposed method over the classic approaches. suggesting the significant potential of sLiDe to enable long-term physiological monitoring using radar.

Ground penetrating radar (GPR) based localization has gained significant recognition in robotics due to its ability to detect stable subsurface features, offering advantages in environments where traditional sensors like cameras and LiDAR may struggle. However, existing methods are primarily focused on small-scale place recognition (PR), leaving the challenges of PR in large-scale maps unaddressed. These challenges include the inherent sparsity of underground features and the variability in underground dielectric constants, which complicate robust localization. In this work, we investigate the geometric relationship between GPR echo sequences and underground scenes, leveraging the robustness of directional features to inform our network design. We introduce learnable Gabor filters for the precise extraction of directional responses, coupled with a direction-aware attention mechanism for effective geometric encoding. To further enhance performance, we incorporate a shift-invariant unit and a multi-scale aggregation strategy to better accommodate variations in dielectric constants. Experiments conducted on public datasets demonstrate that our proposed EDENet not only surpasses existing solutions in terms of PR performance but also offers advantages in model size and computational efficiency.

Recursive techniques have demonstrated excellent computational efficiency in classic remote sensing tasks but are rarely applied in emerging remote sensing tasks in the IoT field, such as bioradar-based remote sensing of vital signs (VS). As a fundamental but important problem, VS sensing lacks an effective state model, making it difficult to implement using recursive techniques for a bioradar system. However, this paper presents an efficient recursive technique (BERT) that benefits remote VS sensing and accelerates computational speed with bioradar. Specifically, the recursive technique is derived from an efficient Markov state model based on the features of bioradar VS and is significantly helpful for algorithm implementation. This technique fills the recursive application gap in the VS sensing task and motivates further exploration of its rationale. Thanks to the strong capability of the recursive technique in computation, the proposed BERT requires less time and demonstrates greater accuracy compared to other methods in simulation results. We further conduct extensive experiments on two real datasets — one comprising 50 children and the other involving 30 adults across various scenarios, and the results show that the BERT algorithm significantly accelerates computation speed, reducing processing time by nearly 41% compared to the state-of-the-art algorithm, while maintaining superior estimation accuracy. Furthermore, this work also offers a fresh perspective on interpreting remote VS sensing for bioradar.

Radar, as a contactless, non-intrusive, weather-and light-independent sensor, works in complementary with other sensors. Radar-based Human Activity Recognition (HAR) has the advantages of privacy preservation and noise robustness. Therefore, it has a tremendous potential for IoT applications. Existing methods are dedicate to finding best radar representations for HAR. However, research on the elimination of non-activity information has not receive sufficient attention. In response to this question, the Crucial Region Search and Feature Discrimination (CRSFD) network has been proposed. It aims to automatically separate features and explicitly establish feature distribution, to accomplish non-activity feature elimination. The CRSFD consists of the Crucial Region Search (CRS) module and the Activity-Associated Feature Discrimination (AAFD) module. The CRS module, with edge protection and importance assessment capabilities, is more suitable for irregular changes in activity features. The AAFD module is customized to automatically and explicitly analyze the distributions of activity and non-activity features. Eventually, activity features are maintained for HAR. Experimental results on a human activity dataset and a human gesture dataset show that the proposed method has superior performance and is robust to noise.

As a promising technology for autonomous driving, localizing ground penetrating radar (LGPR) is a vehicle localization method that relies on prior maps and couples deeply with subsurface features. However, the unique characteristics of GPR data often lead to a significant number of mismatched candidates during localization. Previous learning-based GPR place recognition methods have primarily relied on 2D convolutional neural networks (CNNs), which struggle to effectively capture critical temporal information, limiting further performance improvements. To address this limitation, we propose a spatial-temporal U-shaped network (STU-Net) that leverages 3D convolutional neural networks to simultaneously extract spatial and temporal features from GPR image sequences. Additionally, residual dense blocks (RDBs) are integrated into the network to enable multi-scale feature extraction. Extensive experiments conducted on publicly available datasets demonstrate that our STU-Net achieves state-of-the-art performance, outperforming existing methods with significant improvements.

Ground-penetrating radar (GPR) , as a sensor for mapping and localizing subsurface features, has gained significant attention in robotic localization for complex environments. However, varying weather conditions change the subsurface dielectric constant, which weakens the registration correlation between real-time images and maps, thereby compromising localization stability. We design an image registration framework for localizing ground penetrating radar (LGPR) in multi-weather. Specifically, we first propose an adaptive hyperbolic receptive field that aims to significantly enhance the robust features in the GPR images, while improving both discrimination capability in map and reliability for image registration. Then, an image alignment module is introduced to eliminate the time-delay blurring problem caused by the variation of dielectric constant under multi-weather conditions. The proposed method was evaluated on three distinct datasets (simulated, publicly available, and self-constructed), demonstrating significant improvements in registration and localization performance. The average correlation coefficient on simulated data achieved a enhancement from 0.3796 to 0.7406 compared with baseline, validating the effectiveness of feature enhancement. Furthermore, measured datasets exhibited 15% higher average registration and localization accuracy than baseline. These results demonstrate that the method provides a reliable guarantee for the stable operation of LGPR systems in complex environments. The datasets will be released at: https://github.com/Eazin-bz/dataset_AHFT.git.

In order to adapt to various detection environments on the ground, airborne synthetic aperture radar (SAR) as a remote sensing platform usually arcs along nonlinear trajectories, and the large accumulation angle in the circling process also improves the imaging effect. However, the arc motion demands rigorous control of the flying platform and precise measurement of the motion. In some cases, there is a significant discrepancy between the track recorded by the flight platform and the actual track, which not only affects the imaging effect but also interferes with the positioning and navigation of the platform. This article presents a new method of arc-line SAR positioning and imaging based on a closed-loop structure. The echo history extracted from a 1-D range profile is corrected using an echo-history correction factor (EHCF), which reduces the self-positioning error of the platform caused by the motion measurement device. This allows for accurate positioning of the flying platform and the acquisition of imaging results with superior focusing performance. The effectiveness and stability of the proposed method are proven by simulation and experimental results.

The radar imaging method, based on matched filtering (MF), generates high gratings and sidelobes in sparse aperture data, resulting in artifacts in the radar image. The theory of compressive sensing (CS) has brought a breaking change to radar imaging, and imaging enhancement can be realized by exploiting the sparsity of the target image. However, traditional sparse imaging methods ignore the correlation between scatterers. This leads to difficulties in accurately extracting the target’s shape contour and structural features. Thus, in this paper, a convolutional reweighted model based on structured sparsity features is proposed. Specifically, a dynamically relaxing threshold is achieved through the convolutional reweightedl1norm, promoting the sparsity of clustered structures in radar images. Furthermore, to avoid large-scale matrix inversion, the issue is respectively addressed through the alternating direction method of multipliers (ADMM) joint gradient descent framework and linearization approximation approach. In addition, the priori information of MF is utilized to adaptively update the imaging support set during the iteration process, aiming to reduce the data storage pressure. Finally, a large number of simulation and experimental results confirm the generality of the proposed algorithms for radar data in different frequency bands, as well as their superiority in terms of computational efficiency and image quality.

In view of the complex imaging environment faced by penetrating radar, low-frequency ultra-wideband (UWB) signals are typically employed to achieve better penetration performance and finer distance resolution. Under the conditions of UWB near field, the positioning error of the radar moving platform can have an unignorable impact on the imaging outcome; in a through-the-wall environment, the refraction effect compensation errors brought about by the environmental parameters of wall also severely affect the image quality. At present, existing research can only address one of the platform positioning errors and wall refraction compensation errors, assuming the other to be in an ideal state. However, in practical application, these two types of errors usually coexist, therefore, in this paper, we propose a synchronous compensation method that can jointly estimate radar positioning errors and environmental parameters for compensation, breaking the coupling between the two types of errors and achieving clear imaging of unknown building layouts and internal targets. This method is suitable for Synthetic Aperture Radar (SAR) moving along walls and around buildings. Simulation and measurement experiments have proven the above content.

Measuring vital signs (VS) contained in the echoes is crucial to the analyses of breathing and heartbeat signals using medical radar. Although many advanced signal processing algorithms have been developed for radar-based VS measurement and make some improved progress, existing schemes cannot achieve a good estimation of echo phases modulated by the respiratory and cardiac activities with high accuracy or low computation, and thus resulting in serious performance degradation on the subsequent separation of breathing and heartbeat patterns as well as the assessment of breathing rate (BR), heart rate (HR), and heart rate variability (HRV). In this paper, we propose a simple yet effective method to measure VS for medical radar, named 3M method. Specifically, our method firstly introduces the Markov-Gauss model to obtain the recursive expression of the echo phases carrying VS, and secondly derive a simple observation equation (SOE) to reflect the relationship between the observed signal and VS of radar measurement. Thirdly, the aforementioned Markov-Gauss model and SOE are fused by Kalman filter to measure VS with accurate estimation. The 3M method demonstrates an elegant structure, low complexity and excellent features introduced by Kalman filter. Simulation results show the superiority of 3M over other methods. Then, we conduct extensive experiments with insightful visualizations to validate the effectiveness of the 3M method. Comparative results on different scenarios illustrate that the 3M method not only achieves state-of-the-art VS measurement performance but also expresses robust properties to HRV analysis.

Contactless physiological signal sensing is an emerging technology for routine health monitoring, ambient-assisted living, automotive, search and rescue, and public security. In this work, we propose MetaPhys, a reconfigurable intelligent surface (RIS)-based contactless physiological signal monitoring system that can simultaneously sense multiple persons and their respiratory and cardiac signals. The proposed system localizes the human targets by utilizing RIS to dynamically manipulate the electromagnetic waves in the environment for beamforming and 3-D radar imaging in the spatial dimension. Then, it extracts the respiration and cardiac signals from the sequential echoes in the temporal dimension. The high-resolution information in the four dimensions as “3-D space + 1-D time” makes it comprehensively perceive the targets and environment. The beamforming allows the radiated energy to focus on the thoraxes, which helps reduce interference and distortion caused by stationary reflectors (i.e., clutters, static body parts, and multipath), thereby improving the signal-to-noise ratio and enabling long-range measurements. We also verify the measured data in a real indoor environment, and the experimental results show that the proposed system can accurately monitor the physiological signals of multiple subjects. The root mean square errors of respiratory rate and heart rate at a distance of 3 m are 0.2 breaths per minute and 1.1 beats per minute, respectively.

Dual-function radar and communication (DFRC) has recently drawn significant attention due to its enormous potential. This letter deals with waveform design of DFRC to improve target detectability embedded in clutter environment while guaranteeing the service quality of communication users. Our design objective is to maximize the output signal-to-clutter-plus-noise ratio (SCNR) of multiple-input multiple-output (MIMO) radar, subject to worst-case received symbol errors at communication users. Coordinate descent (CD) as an efficient iteration algorithm is proposed to solve above optimization problem, which splits high-dimensional problem into multiple one-dimensional problem. Furthermore, we introduce Dinkelbach algorithm (DA) to increase rate of convergence, which is an efficient way to reduce complexity. Finally, simulation results are presented to illustrate the effectiveness of the proposed techniques.

It is of great importance to remotely sense trapped victims with radio signals in modern search and rescue after natural disasters like earthquakes, avalanches, building collapses, and so on. Various radio sensors have been developed to date; however, they are hardly deployed to recognize efficiently the vital sign in long-distance, deep-coverage, and multi-subject situations because the back-scattered victim-critical radio signals are really weak and are nearly drowned in the ambient nonstationary noise and clutter. To tackle the formidable difficulty, we present a four-dimensional wideband microwave radar operating at 1.7 GHz to 2.7 GHz and develop a spatio-temporal processing algorithm to fully explore the vital knowledge of victims in the three-dimensional spatial and one-dimensional temporal information. We conducted comprehensive field experiments in real post-disaster environments and demonstrated experimentally that our radio sensor can continuously monitor multiple survivors trapped under mounds of debris in real urban environments. Moreover, we demonstrate that the presented method can achieve a signal-to-noise-and-clutter ratio improvement of more than 20 dB even in the case of deep burial, which enables localizing the victims trapped in the order of ten meters and recognizing the survivors’ vital states. We expect that the presented strategy may open an avenue for future remote life-rescuing and beyond in practical applications.

Radar-based noncontact physiological signals monitoring is meaningful for daily health monitoring, post-disaster rescue, and public security. This paper focuses on the theoretical and experimental study of noncontact respiratory and cardiac motion signals separation by remote sensing using a four-dimensional (4D) imaging radar. To adaptively separate respiratory and cardiac motion patterns, we propose a variational mode separation (VMS) algorithm. VMS is established on optimizing a variational problem to separate different modes. It minimizes the energy overlap of the heartbeat and respiration signals as well as their harmonics with an equality constraint. Both simulation and real scene data results show that the proposed VMS algorithm is suitable for separating the weaker cardiac motion pattern from the strong respiratory motion pattern, restraining the influence of respiration harmonics on the heartbeat component. Furthermore, we have implemented continuous remote monitoring of respiratory rate (RR) and heart rate (HR) by employing the proposed method in a real scene. The results validate the consistency with the reference respiration belt and electrocardiogram (ECG). The root mean square errors (RMSEs) of RR and HR for the remote measurement are 0.13 breaths per minute (brpm) and 1.7 beats per minute (bpm), respectively.

Radar is an extremely valuable sensing technology for detecting moving targets and measuring their range, velocity, and angular positions. When people are monitored at home, radar is more likely to be accepted by end-users, as they already use WiFi, is perceived as privacy-preserving compared to cameras, and does not require user compliance as wearable sensors do. Furthermore, it is not affected by lighting conditions nor requires artificial lights that could cause discomfort in the home environment. So, radar-based human activities classification in the context of assisted living can empower an aging society to live at home independently longer. However, challenges remain as to the formulation of the most effective algorithms for radar-based human activities classification and their validation. To promote the exploration and cross-evaluation of different algorithms, our dataset released in 2019 was used to benchmark various classification approaches. The challenge was open from February 2020 to December 2020. A total of 23 organizations worldwide, forming 12 teams from academia and industry, participated in the inaugural Radar Challenge, and submitted 188 valid entries to the challenge. This paper presents an overview and evaluation of the approaches used for all primary contributions in this inaugural challenge. The proposed algorithms are summarized, and the main parameters affecting their performances are analyzed.

In this letter, we propose a fast superpixel-based clustering algorithm (FSC) for synthetic aperture radar (SAR) image segmentation. First, the SAR image is over-segmented into superpixels by our previously proposed edge-aware superpixel generation method with one iteration merging (ESOM). Second, based on the obtained superpixels, the number of clusters is automatically selected by the density peak (DP) algorithm and knee point method instead of manual specification. Finally, the modified$k$-means clustering with the generalized-likelihood ratio (GLR) dissimilarity is performed on the superpixels to generate the final segmentation result. Experimental results on two real SAR images show that the proposed method outperforms other state-of-the-art methods in terms of both segmentation accuracy and computational efficiency. Moreover, our method is free of clustering parameters and achieves automatic SAR image segmentation.

Autofocusing of multiple-input and multiple-output (MIMO) penetrating radar is a recent developing method to solve the problem of image focusing problem in unknown environment. However, the ergodic search process in the subfocusing process greatly increases the amount of calculation, which limits the application of this method in practice. To solve the problem, we introduce an image-domain-filter-based method. All the compensation and correction are based on image domain, and it avoids calculating the position of the refraction point, which saves much time of computation. In this letter, we prove the image filter can autofocus well in both ground penetrating and wall penetrating scenarios under circumstances with unknown parameters. Both the simulation and the measurement experiments show the method can complete the compensation precisely and quickly and offers better focusing quality.

In recent years, the location and detection of concealed objects and buried targets extensively employed low-frequency MIMO radar, because of its excellent penetration characteristics and high resolution. The environmental parameters are usually unknown, autofocusing method is always used to obtain accurate and clear target images and estimate the environmental parameters at the same time. However, the common disadvantage of the autofocusing methods is the computation burden caused by iterative imaging to estimate the environment parameters. This letter proposes a 3-D image filter method for MIMO array radar, which improve computational efficiency while ensuring the focusing degree and accurate target positioning. Simulation and experimental results show the effectiveness and robustness of the proposed method.

Feature matching refers to the establishment of reliable correspondence between two sets of local features, which is an essential approach in remote sensing applications such as image registration and mosaicking. In this paper, a simple yet effective method, called frame-based locality preservation matching, is proposed for robust remote sensing image matching. We primarily focus on those images pairs that involve large-scale geometric transformations (e.g., extreme zoom). The key idea of our approach is to dig up the frame knowledge, such as the feature orientation and scale implied by common features like SIFT. The frame knowledge is free to obtain, and we find it to be of great significance in feature matching, especially for our focus -- large-scale geometric transformations. The proposed method can easily handle the geometric challenges and high outlier proportions, and significantly improves the performance compared to other state-of-the-art methods.

In this article, we present a content-sensitive superpixel generation method with edge penalty and the contraction–expansion search strategy (EPCES) for synthetic aperture radar (SAR) images. Specifically, the edge information can be obtained by our previously proposed ratio-based edge detector with recurrent guidance filter, which has been proven to be robust to speckle noise and capable of detecting weak edges in low-contrast areas. The content-sensitive superpixel seeds’ initialization method is proposed with respect to the heterogeneous state of the SAR imagery, benefiting from which EPCES can generate an exact number of superpixels set by the user and the fine details can be preserved well. In EPCES, a new dissimilarity with edge penalty is defined to generate the superpixels with better edge adherence. Rather than adopting the conventional clustering method based on local$k$-means, we propose the contraction–expansion search strategy (CES), which explicitly utilizes the continuity information contained in neighboring pixels and enforces the connectivity of the superpixel without any postprocessing step. With the aid of the CES, our proposed method can attain superpixels with low computational cost and high edge adherence. Experimental results on both synthetic and real-world SAR images verify that the proposed method consistently performs favorably against several state-of-the-art methods in terms of both quality and efficiency.

Feature matching is a fundamental technique in remote sensing image processing. This article proposes a new formulation of affine covariant feature matching for remote sensing images, where we suggest matching features by matching two sets of triplets. Compared with previous works, the formulation exploits the whole feature frame rather than the 2-D location to reject outliers. Besides, we also develop a new latent variable model to combine the feature frame and the SIFT ratio values, to enhance the convergence speed and success rate in challenging cases. We evaluate our model on three challenging datasets in terms of both qualitative and quantitative experiments. We also study the robustness to outliers since remote sensing images are typically affected by mismatches. The results demonstrate that the proposed method provides excellent matching performance with satisfying runtime and shows good robustness to outliers.

While traditional edge detectors concentrate on modifying the shape of the window function, we consider the edge detection problem from a new perspective, and an effective recurrent guidance filter is proposed in this letter. The proposed filter is elaborately designed for edge detection tasks and aims to remove the nonedge information including speckle noise and detailed texture and preserve edge information simultaneously. We first filter the image by the proposed filter and a filtered image is obtained. Then, by using the edge detector with the Gaussian-shaped window, which was previously proposed by us and performing the postprocessing method, the edge response is extracted from the filtered image. Both objective and subjective experimental results on simulated and real synthetic aperture radar (SAR) images demonstrate that the edge detector based on the recurrent guidance filter yields better performance than the state-of-the-art edge detectors.

Most of the existing superpixel generation methods are based on local iterative clustering. However, such methods have the following shortcomings: 1) these methods require several iterations and the number of iterations is difficult to determine and 2) the generated superpixel lacks explicit connectivity without a postprocessing step. Aiming to overcome the limitations, we propose an edge-aware superpixel generation with one iteration merging (ESOM) for synthetic aperture radar (SAR) imagery. In specific, we introduce a ratio-based edge detector with a Gaussian-shaped window to extract the edge information and an edge-aware dissimilarity is defined. Then, a new merging method termed as one iteration merging is proposed, which leverages the continuity of the adjacent pixels and ensures the connectivity of superpixel. Furthermore, instead of iterative clustering, the one iteration merging is achieved in only one iteration without determining the number of iterations and hence efficient in computation. Experiments on two real SAR images demonstrate that the proposed method yields substantially better performance than some state-of-the-art methods.

Limited by the total length, the total number of the antenna units as well as their topology, the radar images always suffered from the sidelobe/grating lobe which severely impacts the quality of the radar images. In this article, a convolutional neural network (CNN)-based radar image-enhancing method is proposed. Using the original radar images as the input samples and using their corresponding ideal radar images with no sidelobe/grating lobe as the label to train the CNN. A well-trained CNN can suppress the sidelobe/grating lobe in the radar images. The structure of the specific CNN, the generation methods of the samples and the labels, the training procedure of the CNN, as well as some other detailed implementation strategies are specifically illustrated in this article. The proposed method is utilized to suppress the sidelobe/grating lobe in both the simulated and real recorded radar images. Compared to other existing methods, the proposed method is with better sidelobe/grating lobe suppressing performance and better robustness.

The fundamental difficulties in the supervised deep learning algorithm are obtaining large-scale labeled data and generalizing the trained model to a new environment. In this letter, we propose an unsupervised domain adaption method for human activity classification using micro-Doppler signatures. We study on how to classify micro-Doppler signatures in a new domain using only labeled samples from a different domain, mainly focus on simulation-to-real-world deep domain adaptation. First, we use motion capture (MOCAP) database to generate simulated micro-Doppler data to train the convolutional neural network (CNN). Then, considering the difference between simulation and real-world domain distributions, we introduce a domain discriminator to pit against the feature extractor part of the CNN. Through this adversarial process, like the generative adversarial network, the CNN trained on the simulation domain is able to generalize to the real-world domain. Experiment results show that the proposed method achieves over 84.02% accuracy in real-world micro-Doppler classification, which outperforms nearly 16% in CNN trained on the annotated simulation without domain adaptation and performs better than the existing domain adaptation methods.

Deep neural networks have shown promise in the radar-based human activity analysis application. Different from existing deep learning models that take either micro-Doppler spectrograms or range profiles as their input, the proposed method can process micromotion signatures in a 3-D way. In this letter, we first transform radar echoes into range-Doppler (RD) time points and then directly process the point sets via a designed 3-D network called the RD PointNet. In fact, our point model is a discrete representation of the motion trajectory. Through this quantitative model, we can use the 3-D network to simultaneously capture human motion profiles and temporal variations. The motion capture simulations and ultrawideband radar measurements show that the proposed framework can achieve superior classification accuracy and noise robustness when compared with image-based methods.

This letter is devoted to the layout reconstruction via the ultra-wideband (UWB) through-wall imaging radar under one single observation and simultaneously takes account of the real-time human indication. In the proposed framework, layout reconstruction is taken as the preprocessing, where a coherent processing interval consisting of several successive received echoes in the initial stage is first employed to construct a range-Doppler (RD) spectrum. Then, in the RD spectrum, a series of selected discrete Doppler frequency signals is used to form Doppler back projection (BP) images. Finally, in the Doppler BP image stack, we design a 3-D constant false alarm rate detector to extract the building layout. Once completed, the achieved layout as auxiliary information is fused with the simultaneous human indication. Through-wall experiments show that the proposed method can effectively extract the covered layout of multiple walls under one single view and accordingly provide strong support for the concealed human sensing.

The work presented in this letter focuses on the time-variant radio frequency interference (RFI) issue in the low-frequency ultra-wideband (UWB) radar. Different from many previous studies, we first analyze the characteristics of RFIs and scattered echoes in the slow-time dimension and take advantage of overlapped short-time Fourier transform to adapt to the time-variant RFIs and update the frequency Doppler spectrum. Then, in the frequency Doppler spectrum, we adopt the minimum statistic combined with 1-D cell-averaging constant false alarm rate to estimate and separate the RFI power spectrum from the scattered echoes based on their differences. Finally, to mitigate the estimated RFIs, a suboptimal filter controlled by the defined entropy of range profiles after math filtering is designed. Employing a UWB radar, different experiments were conducted, and results verify the proposed method.

In this letter, we propose an isophase-based method for human life signal extraction using a stepped-frequency continuous-wave through-the-wall radar. Isophase is the line connected by points of the same phase value in the slow time range map and is a key feature to distinguish a stationary human from mechanical vibrating targets. First, the isophase is introduced and associated linearly with human life signal, and the linear coefficient is greater than 1. Then the limitation of the isophase is discussed. Finally, the isophase is extracted by the edge detection method. The isophase expands the micromotion range and therefore robust to noise. Due to the sudden change from λ to -λ, the isophase can remove outliers effectively. The simulation results verify its advantages over the amplitude peak locus. The experiments done with a stationary human and a rotating fan behind a wall show that the proposed method can extract human life signal effectively and precisely.

Human target indication and detection are widely used in military and civil applications. For stationary human targets behind an obstacle, the micro-motion caused by vital sign results in a fairly low signal-to-clutter-and-noise (SCNR). This study presents an iterative singular vector decomposition (ISVD) based micro-motion indication method for detecting human targets with through-the-wall radar. First, singular vector decomposition (SVD) with stepped frequency signal is analyzed, and then iterative SVD is proposed to indication micro-motion human targets. Through this iterative process, the residual clutter and noise are effectively removed. Simulation and experimental results indicate that the ISVD based method is robust to noise and improve SCNR remarkably.

Human target detection and tracking have great potential in military, safety, security and entertainment applications. In this paper, a complete processing procedure is proposed for human tracking in foliage-penetration environment by multistatic radar. It consists of five main steps, including clutter suppression, target detection, measurement estimation, target localization and target tracking. Exponential average background subtraction is applied for clutter suppression. Ordered statistics constant false alarm rate (OS-CFAR) detector is used to detect targets in the range profile. The range measurement estimation is realized by a window filter. The S-D assignment algorithm is introduced for multi-target localization. Target tracking is realized by a Kalman filter based multi-target tracking (MTT) system. The experimental results verify the effectiveness of the proposed human tracking procedure.

Multistatic radar system has a great potential for human detection and tracking for its fine localization precision, wide coverage and good observability. T-RN(one transmitting antenna and N receiving antennas) radar system is the most popular multistatic radar structure. In this paper, the localization performance of T-RNradar is evaluated using Cramér-Rao lower bound (CRLB). The localization precision of different radar topology types and different numbers of receiving antennas are compared. The localization performance is also evaluated by experimental data. The results reveal that more separated receiving antenna configurations and more numbers of receiving antennas have better localization precision.

In through-wall indication of a moving human target in enclosed structures, a shadow effect because of the human target blocking parts from illumination on the back wall will emerge, referred to as a “ghost” in indication results. The shadow ghost moves as the human target does, which makes causal change detection (CD) invalid to separate them. To mitigate the shadow ghost, we analyze its differences from the moving human target. Based on the difference that the illumination is only blocked in partial channels of the multiple-input–multiple-output (MIMO) array while target echoes exist in most channels and the fact that shadow ghosts overlap more between successive indication results than the imaged targets as a result of their larger size, we proposed a mitigation method including a coherence factor and noncausal CD processing. Through-wall experiments via a MIMO through-wall imaging radar validate the proposed method.

Soon after its innovation, compressive sensing (CS) was rapidly applied to radar imaging. However, the huge computational complexity and the memory requirements have become the bottlenecks in its widespread applications to large-scale and real-time radar imaging. In this paper, two novel methods based on fast Gaussian gridding nonuniform fast Fourier transform are proposed to speed up CS radar imaging and reduce the memory requirement. By using the proposed methods, the application of CS imaging method can be extended to large-scale and real-time radar imaging with high reconstructing efficiency and small memory requirement. Theoretical analysis and numerical results from the aspects of accuracy, efficiency, and memory requirement validate the proposed methods. Simulation and real data imaging results by spectral projection gradient ℓ1-norm method are given to further demonstrate the efficiency of the proposed methods.

Through-wall indication of human targets is highly desired in many applications. Generally, human targets behind wall are noncooperative, and rare prior knowledge about the circumstance behind wall could be available. Thus, it requires the ability to indicate human targets with different motions from clutters. To investigate this problem, we first examine the conventional time-domain indication methods, and find that their performances are controlled by the historical pulse number adopted to estimate background, which corresponds to the tap-length from the angle of filter. Then, based on an intermittent mode of human target echoes, we define the optimum tap-length as the shortest tap-length that makes the filter output signal-to-clutter-and-noise ratio reach maximum and develop an adaptive indication method with a gradient tap-length control scheme to search the optimum tap-length. Finally, through-wall experiments with an impulse through-wall radar demonstrate that the proposed method can obtain a good adaptive indication performance on human target with different motions.

Compressed sensing (CS) based radar imaging requires the use of a mathematical model of the observation process. Inaccuracies in the observation model may cause defocusing in the reconstructed images. In the observation process, the observation positions are usually not known perfectly. Imperfect knowledge of the observation positions is a major source of model errors in imaging. In this paper, a method is proposed to compensate the observation position errors in CS-based radar imaging. Instead of treating the observation-position-induced model errors as phase errors in the data, the proposed method can determine the observation position errors as part of the imaging process. It uses an iterative algorithm, which cycles through steps of target reconstruction and observation position error estimation and compensation. The proposed method can estimate the observation position errors accurately, and the reconstruction quality of the target images can be improved significantly. Simulation results and experimental results from rail-mounted radar and airborne synthetic aperture radar are presented to show the effectiveness of the proposed method.

This paper presents a sparse multiple-input and multiple-output (MIMO) array and sparse frequency ground-penetrating radar (GPR) imaging scheme based on compressed sensing (CS). Since the targets of interest for GPR are usually sparse, the number of the MIMO array elements and frequencies can be reduced using CS theory. Thus, the system complexity and data acquisition time can be reduced accordingly. Considering the serious clutter in forward-looking GPR, we propose two methods for the CS reconstruction in clutter environment. The first one is a clutter suppression preprocessing method, which can effectively suppress the azimuth clutter and short range clutter outside the reconstruction region and significantly improve the reconstruction result. The second one is to determine the regularization parameter for the CS reconstruction in clutter environment. We refer to this reconstruction process as basis pursuit declutter. The proposed imaging scheme can produce pointlike and less cluttered images of sparse targets using fewer array elements and frequencies. Results from simulated data, trihedral reflector, and real buried land mine experimental data are presented to show the validity of the proposed methods. The experimental data are acquired by the vehicle-mounted stepped-frequency forward-looking ground-penetrating virtual aperture radar, which is designed and developed by the National University of Defense Technology.

Stepped-frequency waveforms can achieve an ultrawide bandwidth by using a sequence of single-frequency pulses. The advantages of stepped-frequency waveforms are low hardware requirements and high resolution. However, the stepped-frequency waveform requires a long time period to transmit the signals, which limits its application in synthetic aperture radar (SAR). The available imaging range width is usually very narrow, unless the range and azimuth resolutions are both decreased. In this paper, a random-frequency SAR imaging scheme based on compressed sensing is proposed. If the targets are sparse or compressible, it is sufficient to transmit only a small number of random frequencies to reconstruct the image of the targets. This means that the limitations of the stepped-frequency technique for SAR can be overcome. The available imaging range width can be enlarged significantly, while the range and azimuth resolutions are both maintained. Random undersampling is very easy to implement for both range and azimuth dimensions, and no new hardware components are needed. Simulation and experimental results are presented to demonstrate the validity of the proposed method.

The compressed sensing (CS) synthetic aperture radar (SAR) imaging scheme can use random undersampled data to reconstruct images of sparse or compressible targets. However, compared to Nyquist sampling, the cost of the CS imaging scheme is the long reconstruction time, particularly for the conventional reconstruction strategy, which reconstructs the whole scene in one process. It also needs a large memory to access the sensing matrix used for reconstruction. In this paper, a segmented reconstruction strategy for the CS SAR imaging scheme is proposed. The whole scene is split into a set of small subscenes, so that the reconstruction time can be reduced significantly. The proposed method also needs much less memory for computation than the conventional method. In this proposed method, the range profiles are reconstructed first, and then, the range profiles can be split into subpatches. Subscenes can be reconstructed by using the subpatch data, and the whole scene can be obtained by combining the reconstructed subscenes. Simulation and experimental results are shown to demonstrate the validity of the proposed method.

A system with 3-D imaging capability can be implemented by using a frequency modulated continuous wave (FMCW) radar which synthesizes a two-dimensional (2-D) planar aperture. A millimeter-wave FMCW three-dimensional (3-D) imaging system can be used for the detection of concealed weapons and contrabands at airports or other security checkpoints, since millimeter-wave can readily penetrate common clothing material. A 3-D image can be formed by coherently integrating the backscatter data over the measured frequency bandwidth and the two spatial coordinates of the 2-D synthetic aperture. This paper presents a 3-D imaging algorithm for near field FMCW radar. This algorithm is an extension of the 2-D range migration algorithm (RMA). We derive the formulation in detail by using the principle of stationary phase (POSP). A 3-D version Stolt interpolation is used in this algorithm. Accurate image reconstruction and high computational efficiency of this algorithm are demonstrated through simulation results.

An extended two-step focusing approach (ETSFA) for processing the squinted spotlight synthetic aperture radar (SAR) data is proposed in this paper. The effect of the squint angle on the azimuth coarse focusing is analyzed and discussed. Based on the analysis results, a nonlinear shift preprocessing method is introduced, which can completely remove the squint angle impacts on the azimuth coarse focusing. Furthermore, based on the squinted spotlight SAR imaging model and the preprocessed echo data, derivations of the azimuth coarse focusing with the deramping-based technique and precise focusing with the modified Stolt-based technique are carried out in detail. Moreover, to produce an acceptable image by the proposed ETSFA for high-resolution ($<$3 m) squinted spotlight SAR with large scene, a subscene processing method is introduced. The experimental results on simulated data prove the validity of the whole analysis and the proposed methods.

In this paper, an extended nonlinear chirp scaling (ENLCS) algorithm for focusing synthetic aperture radar data acquired at high resolution and highly squint angle is proposed. The whole processing of the ENLCS consists of the following three steps. First, a linear range walk correction is used to remove the linear component of target range cell migration (RCM) and to mitigate the range-azimuth coupling of the 2-D spectrum. Second, a bulk second range compression (SRC) is performed in the 2-D frequency domain for compensating the residual RCM, SRC term, and higher order range-azimuth coupling terms. Third, a modified azimuth NLCS (ANLCS) operation is applied to equalize the azimuth frequency modulation rate for azimuth compression. By adopting higher order approximation processing and by properly selecting the scaling coefficients, the proposed modified ANLCS operation has better accuracy and little image misregistration. The overall focusing procedure of the ENLCS algorithm only involves fast Fourier transform and complex multiplication, which means easier implementation and higher efficiency. The experimental results with simulated data prove the effectiveness of the proposed algorithm.

This paper presents a synthetic aperture radar (SAR) imaging system using a stepped frequency waveform. The main problem in stepped frequency SAR is the range difference in one sequence of pulses. This paper analyzes the influence of range difference in one sequence of pulses in detail and proposes a method to compensate this influence in the Doppler domain. A Stolt interpolation for the stepped frequency signal is used to focus the data accurately. The parameters for the stepped frequency SAR are analyzed, and some criteria are proposed for system design. Finally, simulation and experimental results are presented to demonstrate the validity of the proposed method. The experimental data are collected by using a stepped frequency radar mounted on a rail.

We propose here a new approach for synthetic aperture radar (SAR) imaging of ground moving targets. The unique characteristic of this approach is that range curvature (i.e., quadratic range migration) can be corrected by a simple processing step. A keystone transform is used to correct range walk (i.e., linear range migration) for all targets without knowing their velocities, and the range curvature is corrected in the range-Doppler domain. The advantage of this approach is that it is simple to implement and can correct range curvature for all targets in one processing step, so that it is computationally efficient. Simulation and experimental SAR data processing results are presented to demonstrate the validity of the proposed approach.

Phase adjustment by contrast enhancement (PACE) is an autofocus algorithm that is capable of performance that is unattainable by conventional techniques. It is a nonparametric method that requires no constraints on the type of phase error to be measured. The algorithm does not require special data culling techniques or the presence of isolated scatterers. However, the drawback of PACE algorithm is that the number of estimated variables is very large; it leads to a long computational time. The azimuth sampling frequency is commonly much bigger than the bandwidth of phase error in SAR image so that we can estimate part of the phase error variables and then obtain the whole variables by interpolation; this induces the interpolated phase adjustment by contrast enhancement (IPACE) algorithm. The IPACE algorithm can remarkably reduce the computational time while maintaining the accuracy. This letter has derived the detailed processing of IPACE, and the results of the experiments using real SAR data are presented to show the validity of the proposed algorithm.