Search

Micro-Doppler signals analysis has been emerging as an important topic in target identification, and relative research has been focusing on features extraction and separation of the radar signals. As a time-frequency representation, the Hilbert-Huang transform (HHT) could extract the accurate instantaneous micro-Doppler signature from the radar signals by empirical mode decomposition and Hilbert transform. However, HHT has the shortcoming that it cannot decompose the signals with close-frequency components. To solve this problem, an innovative decomposition method for multicomponent micro-Doppler signals based on Hilbert–Huang transform and analytical mode decomposition (HHT-AMD) is proposed. In this method, the multicomponent micro-Doppler signals are firstly decomposed by empirical mode decomposition, and the decomposed signal components are transformed by Hilbert transform to get the Hilbert-Huang spectrum and marginal spectrum. Through the spectrum processing, we get the frequency distribution of each signal component. The next step is to judge whether there exists frequency aliasing in each signal component. If there is aliasing, the AMD method is used to decompose the signal until all the decomposed signals are mono-component signals. Evaluation considerations are covered with numerical simulations and experiments on measured radar data. The results demonstrate that compared with conventional HHT, the proposed method yields accurate decomposition for multicomponent micro-Doppler signals and improves the robustness of decomposition. The method presented here can also be applied in various settings of non-stationary signal analysis and filtering.

Radar has great potential in military and civilian areas, including automobile anti-collision, battlefield surveillance, etc., due to its high penetration and all-weather capability. On the basis of traditional targets detection, targets classification can be realized. In this paper, a comparison of targets classification between deep learning (Deep Convolutional Neural Networks (DCNNs)) and conventional supervised learning methods (Support Vector Machine (SVM), Naive Bayes (NB) and SVM-Bayes fusion algorithm) has been made. Furthermore, several factors affecting the accuracy of classifying targets including SNR, decrease of samples, have been researched and discussed. We employ a K-band Doppler radar to acquire the raw signal due to its stationary clutter-rejection, movement detection ability and short wavelength. Then Short-time Fourier Transform (STFT) is applied to the raw signal to characterize micro-Doppler signatures which is the fundament of the classification process. We adopt the DCNNs to deal with the spectrograms directly, while features have been designed and extracted for classification with conventional supervised learning methods. It is shown that the DCNN can achieve average accuracy approximately 99.4% followed by SVM-Bayes fusion algorithm reaching around 95.8%, while the accuracy for SVM and NB is about 94.4% and 91% respectively.

Radar-based activity recognition is a problem that has been of great interest due to applications such as border control and security, pedestrian identification for automotive safety, and remote health monitoring. This paper seeks to show the efficacy of micro-Doppler analysis to distinguish even those gaits whose micro-Doppler signatures are not visually distinguishable. Moreover, a three-layer, deep convolutional autoencoder (CAE) is proposed, which utilizes unsupervised pretraining to initialize the weights in the subsequent convolutional layers. This architecture is shown to be more effective than other deep learning architectures, such as convolutional neural networks and autoencoders, as well as conventional classifiers employing predefined features, such as support vector machines (SVM), random forest, and extreme gradient boosting. Results show the performance of the proposed deep CAE yields a correct classification rate of 94.2% for micro-Doppler signatures of 12 different human activities measured indoors using a 4 GHz continuous wave radar—17.3% improvement over SVM.

Radar sensors are utilized for detection and classification purposes in various applications. In order to use deep learning techniques, lots of training data are required. Accordingly, lots of measurements and labelling tasks are then needed. For the purpose of pre-training or examining first ideas before bringing them into reality, synthetic radar data are of great help. In this paper, a workflow for automatically generating radar data of human gestures is presented, starting with creating the desired animations until synthesizing radar data and getting the final required dataset. The dataset could then be used for training deep learning models. A classification scenario applying this workflow is also introduced., acceptedVersion

The authors acknowledge the funding received from the Science and Technology Facilities Council which has supported this work under grant ST/N006569/1. This paper presents the hardware design of a coherent 24 GHz radar system developed at the University of St Andrews to obtain micro-Doppler data. The system is based on the Analog Devices EV-RADAR-MMIC2 evaluation board which is based around a chipset of three integrated circuits: a two channel transmitter, a four channel receiver and a fractional-N frequency synthesizer. The evaluation board is combined with a number of other components to enable coherent operation with a PC-based data acquisition card and to boost the output power to increase the operational range. Three identical custom-made smooth-walled conical horn antennas for transmit and co- and cross-polar receive signals were designed and built for the radar system. It is shown that the performance of these high gain (24.5 dBi) antennas agrees extremely well with the design simulations. Finally, field trial results comprising human, bird and drone micro-Doppler data are shown to validate the system performance. Publisher PDF

A simplified polarimetric model for radar signal scattering on wind turbines (WT) is proposed, which can be used for interpretation of observation in time and Doppler domains. This model uses representation of real WT as slowly rotating linear wire structures. The earlier proposed model has been extended to more general cases of WT with arbitrary orientation and full polarimetric observations in mono- and bi-static cases, which can be used for analysis of low-frequency signals scattering. It gives a possibility to estimate the influence of scattered on WT electromagnetic waves not only on radars but also on communication links.

Contemporary surveillance systems mainly use video cameras as their primary sensor. However, video cameras possess fundamental deficiencies, such as the inability to handle low-light environments, poor weather conditions, and concealing clothing. In contrast, radar devices are able to sense in pitch-dark environments and to see through walls. In this paper, we investigate the use of micro-Doppler (MD) signatures retrieved from a low-power radar device to identify a set of persons based on their gait characteristics. To that end, we propose a robust feature learning approach based on deep convolutional neural networks. Given that we aim at providing a solution for a real-world problem, people are allowed to walk around freely in two different rooms. In this setting, the IDentification with Radar data data set is constructed and published, consisting of 150 min of annotated MD data equally spread over five targets. Through experiments, we investigate the effectiveness of both the Doppler and time dimension, showing that our approach achieves a classification error rate of 24.70% on the validation set and 21.54% on the test set for the five targets used. When experimenting with larger time windows, we are able to further lower the error rate.

Video cameras are arguably the world’s most used sensors for surveillance systems. They give a highly detailed representation of a situation that is easily interpreted by both humans and computers. However, these representations can lose part of their representational value when being recorded in less than ideal circumstances. Bad weather conditions, low-light illumination or concealing objects can make the representation more opaque. A radar sensor is a potential solution for these situations, since it is unaffected by the light intensity and can sense through most concealing objects. In this paper, we investigate the performance of a structured inference network on data of a low-power radar device. A structured inference network applies automated feature extraction by creating a latent space out of which the observations can be reconstructed. A classification model can then be trained on this latent space. This methodology allows us to perform experiments for both person identification and action recognition, resulting in competitive error rates ranging from 0% to 6.5% for actions recognition and 10% to 12% for person identification. Furthermore, the possibility of a radar sensor being used as a complement to a camera sensor is investigated.

Continuous in-home monitoring of older adults living alone aims to improve their quality of life and independence, by detecting early signs of illness and functional decline or emergency conditions. To meet requirements for technology acceptance by seniors (unobtrusiveness, non-intrusiveness, privacy-preservation), this study presents and discusses a new smart sensor system for the detection of abnormalities during daily activities, based on ultra-wideband radar providing rich, not privacy-sensitive, information useful for sensing both cardiorespiratory and body movements, regardless of ambient lighting conditions and physical obstructions (through-wall sensing). The radar sensing is a very promising technology, enabling the measurement of vital signs and body movements at a distance, and thus meeting both requirements of unobtrusiveness and accuracy. In particular, impulse-radio ultra-wideband radar has attracted considerable attention in recent years thanks to many properties that make it useful for assisted living purposes. The proposed sensing system, evaluated in meaningful assisted living scenarios by involving 30 participants, exhibited the ability to detect vital signs, to discriminate among dangerous situations and activities of daily living, and to accommodate individual physical characteristics and habits. The reported results show that vital signs can be detected also while carrying out daily activities or after a fall event (post-fall phase), with accuracy varying according to the level of movements, reaching up to 95% and 91% in detecting respiration and heart rates, respectively. Similarly, good results were achieved in fall detection by using the micro-motion signature and unsupervised learning, with sensitivity and specificity greater than 97% and 90%, respectively.

Classification modes aimed for slow events, such as human gait, using micro-Doppler measurements are long, and therefore not suitable for use during time critical radar operation. When these modes are interruptible by other radar modes they might become more acceptable. In this paper, sparse signal processing, in particular the SL0-algorithm, is investigated for interpolating interrupted radar measurements which subsequently can be used for classification of human gaits. The performance of a k-NN classifier with PCA-based feature extraction was improved significantly when the data were interpolated as compared to using the interrupted data without interpolation.

The authors acknowledge the funding received from the Science and Technology Facilities Council which has supported this work under grant ST/N006569/1. This paper discusses the micro-Doppler signatures of small UAVs obtained from a millimeter-wave radar system. At first, simulation results are shown to demonstrate the theoretical concept. It is illustrated that whilst the propeller rotation rate of the small UAVs is quite high, millimeter-wave radar systems are capable of capturing the full micro-Doppler spread. Measurements of small UAVs have been performed with both CW and FMCW radars operating at 94 GHz. The CW radar was used for obtaining micro-Doppler signatures of individual propellers. The field test data of a flying small UAV was collected with the FMCW radar and was processed to extract micro-Doppler signatures. The high fidelity results clearly reveal features such as blade flashes and propeller rotation modulation lines which can be used to classify targets. This work confirms that millimeter-wave radar is suitable for the detection and classification of small UAVs at usefully long ranges. Publisher PDF

Son zamanlarda, derin yapay sinir ağları (DNN), sentetik açıklılı radar (SAR) veya mikro-Doppler tabanlı otomatik hedef tanıma gibi RF sinyal sınıflandırma uygulamalarında araştırma konusu olmuştur. Bununla birlikte, radar verisi elde etmenin masraflı ve zor olması, elde edilebilecek veri sayısını kısıtlamaktadır. DNN'ler yüksek varyanslı modeller olmaları sebebiyle aşırı öğrenmeye yatkıdırlar. Dolayısıyla iyi bir genelleme öğrenebilmek için fazla miktarda etiketlenmiş veriye ihtiyaç duymaktadırlar. Bu tez çalışmasında, ilk olarak 4 GHz sürekli dalga radarı 12 farklı ancak birbirine benzer insan hareketi için bir veri seti oluşturulmuştur. Sonrasında literatürde az sayıda veriyle DNN eğitmek için önerilen iki temel teknik kıyaslanmıştır: Gözetimsiz ön-eğitim ve aktarımlı öğrenme. Gözetimsiz ön-eğitim bir konvolüsyonel otokodlayıcı (CAE) yapısı üzerinden, aktarımlı öğrenme ise ImageNet üzerine eğitilmiş popüler CNN mimarileri (VGGNet, GoogleNet and ResNet) üzerinden gerçekleştirilmiştir. Buna ek olarak, aktarımlı öğrenmenin benzer transfer alanı üzerinden uygulanabilmesi için, farklılaştırılmış radar mikro-Doppler simülasyonları oluşturan yenilikçi bir metot önerilmiştir. Tüm sonuçlar, el işçiliği ile oluşturulmuş özniteliklerin çıkartılarak bir SVM sınıflandırıcısının eğitildiği geleneksel yöntem ile kıyaslanmıştır. Tezin sonunda, önerilen artçıl aktarımlı öğrenme ağı DivNet'in, diğer DNN mimarileri arasında en yüksek sınıflandırma doğruluğunu sağladığı ve gürültüye karşı gürbüz olduğu gösterilmiştir., Recently, deep neural networks (DNNs) have been the subject of intense research for the classification of radio frequency (RF) signals, such as synthetic aperture radar (SAR) imagery or micro-Doppler based automatic target recognition. However, a fundamental challenge is the typically small amount of data available due to the high costs and resources required for measurements. Deep neural networks (DNNs), however, require large-scale labeled datasets to prevent overfitting while having good generalization. In this thesis, the efficacy of two neural network initialization techniques - unsupervised pre-training and transfer learning - for dealing with training DNNs on small data sets are compared. Unsupervised pre-training is implemented through the design of a convolutional autoencoder (CAE), while transfer learning from two popular CNN architectures (VGGNet, GoogleNet and ResNet). Also, a novel method for generating diversified radar micro-Doppler signatures using Kinect-based motion capture simulations is proposed as a training database for DNNs exploiting transfer learning. Results have also been compared with the conventional approach where handcrafted features are employed to train an SVM classifier. In particular, it is shown that the proposed residual transfer learning network, DivNet, offers the highest classification accuracy among other DNN architectures and, is robust to image noise.

In inverse synthetic aperture radar (ISAR) imaging, time-frequency analysis is the basic method for processing echo signals, which are reflected by the results of time-frequency analysis as each component changes over time. In the time-frequency map, a target&rsquo, s rigid body components will appear as a series of single-frequency signals in the low-frequency region, and the micro-Doppler components generated by the target&rsquo, s moving parts will be distributed in the high-frequency region with obvious frequency modulation. Among various time-frequency analysis methods, S-transform is especially suitable for analyzing these radar echo signals with micro-Doppler (m-D) components because of its multiresolution characteristics. In this paper, S-transform and the corresponding synchrosqueezing method are used to analyze the ISAR echo signal and perform imaging. Synchrosqueezing is a post-processing method for the time-frequency analysis result, which could retain most merits of S-transform while significantly improving the readability of the S-transformation result. The results of various simulations and actual data will show that S-transform is highly matched with the echo signal for ISAR imaging: the better frequency-domain resolution at low frequencies can concentrate the energy of the rigid body components in the low-frequency region, and better time resolution at high frequencies can better describe the transformation of the m-D component over time. The combination with synchrosqueezing also significantly improves the effect of time-frequency analysis and final imaging, and alleviates the shortcomings of the original S-transform. These results will be able to play a role in subsequent work like feature extraction and parameter estimation.

Under the same principle, laser radar could be more sensitive to the micro-Doppler (m-D) effect due to its wave length, as the characteristic of multi-resolution, S transform could reduce the influence of the micro-Doppler component and enhance the imaging effect. This paper presents a method for micro-Doppler feature extraction in Inverse Synthetic Aperture Imaging Laser Radar (ISAIL) imaging. It is accessible and comprehensive, applying Random Sample Consensus (RANSAC) for the separation and reconstruction of micro-Doppler and rigid body signals. Experiments show that the method can effectively remove the micro-Doppler information and obtain a clear target distance-instantaneous Doppler image.

Micro-Doppler signature produced by micro-motion contains movement and structure information which is useful for radar classification and recognition. In order to recognize the wheeled vehicle and the tracked vehicle in ground battle, the motion dynamics and echo models of the scattering point on wheeled vehicle and tracked vehicle are set up respectively. Then the micro-Doppler features of them are quantitatively analyzed. Furthermore, the smoothed pseudo Winger-Ville distribution method is proposed for extracting micro-Doppler modulation. Finally, simulation results demonstrate the correctness of the theoretic analysis and the validity of the extraction method.

Falls are a major cause of fatal and nonfatal injuries in people aged 65 years and older. Radar has the potential to become one of the leading technologies for fall detection, thereby enabling the elderly to live independently. Existing techniques for fall detection using radar are based on manual feature extraction and require significant parameter tuning in order to provide successful detections. In this paper, we employ principal component analysis for fall detection, wherein eigen images of observed motions are employed for classification. Using real data, we demonstrate that the PCA based technique provides performance improvement over the conventional feature extraction methods. 2016 SPIE. Scopus

This data-set contains signals from scaled replicas of ballistic targets with micro-motions acquired in a controlled indoor environment with a 24GHz CW radar. This work was supported by the Engineering and Physical Sciences Research Council (EPSRC) Grant number EP/K014307/1 and by the Centre for Defence Enterprise under the contract CDE36521. The data which underpins this paper is subject to a confidentiality agreement with one of the collaborators - as such the data cannot be made openly available. Enquiries about this restriction can be submitted to the contact email listed on this page.