@article{manjur_senyurek_kalski_gupta_skarke_gurbuz_2025, title={Automated Detection of Seafloor Gas Seeps in Multibeam Echosounder Data With an Attention-Guided Convolutional Neural Network}, url={https://doi.org/10.1109/JSTARS.2025.3535234}, DOI={10.1109/JSTARS.2025.3535234}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, author={Manjur, Sultan Mohammad and Senyurek, Volkan and Kalski, Ramon and Gupta, Surabhi and Skarke, Adam and Gurbuz, Ali C.}, year={2025} }
@article{alam_farhad_al-qwider_owfi_koosha_mastronarde_afghah_marojevic_kurum_gurbuz_2024, title={A Physical Testbed and Open Dataset for Passive Sensing and Wireless Communication Spectrum Coexistence}, url={https://doi.org/10.1109/ACCESS.2024.3453774}, DOI={10.1109/ACCESS.2024.3453774}, journal={IEEE Access}, author={Alam, Ahmed Manavi and Farhad, Md Mehedi and Al-Qwider, Walaa and Owfi, Ali and Koosha, Mohammad and Mastronarde, Nicholas and Afghah, Fatemeh and Marojevic, Vuk and Kurum, Mehmet and Gurbuz, Ali C.}, year={2024} }
@article{king_asaduzzaman towfiq_gurbuz_cetiner_2024, title={Beam Coefficient Prediction for Antenna Arrays Using Physics-Aware Convolutional Neural Networks}, volume={12}, ISSN={["2169-3536"]}, url={https://doi.org/10.1109/ACCESS.2024.3491828}, DOI={10.1109/ACCESS.2024.3491828}, journal={IEEE ACCESS}, author={King, Glendyn D. and Asaduzzaman Towfiq, Md and Gurbuz, Ali C. and Cetiner, Bedri A.}, year={2024}, pages={176908–176919} }
@article{nabi_senyurek_kurum_gurbuz_2024, title={Best Linear Unbiased Estimators for Fusion of Multiple CYGNSS Soil Moisture Products}, url={https://doi.org/10.1109/JSTARS.2024.3443100}, DOI={10.1109/JSTARS.2024.3443100}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, author={Nabi, M M and Senyurek, Volkan and Kurum, Mehmet and Gurbuz, Ali Cafer}, year={2024} }
@article{alaba_gurbuz_ball_2024, title={Emerging Trends in Autonomous Vehicle Perception: Multimodal Fusion for 3D Object Detection}, url={https://www.mdpi.com/2032-6653/15/1/20}, DOI={10.3390/wevj15010020}, abstractNote={The pursuit of autonomous driving relies on developing perception systems capable of making accurate, robust, and rapid decisions to interpret the driving environment effectively. Object detection is crucial for understanding the environment at these systems’ core. While 2D object detection and classification have advanced significantly with the advent of deep learning (DL) in computer vision (CV) applications, they fall short in providing essential depth information, a key element in comprehending driving environments. Consequently, 3D object detection becomes a cornerstone for autonomous driving and robotics, offering precise estimations of object locations and enhancing environmental comprehension. The CV community’s growing interest in 3D object detection is fueled by the evolution of DL models, including Convolutional Neural Networks (CNNs) and Transformer networks. Despite these advancements, challenges such as varying object scales, limited 3D sensor data, and occlusions persist in 3D object detection. To address these challenges, researchers are exploring multimodal techniques that combine information from multiple sensors, such as cameras, radar, and LiDAR, to enhance the performance of perception systems. This survey provides an exhaustive review of multimodal fusion-based 3D object detection methods, focusing on CNN and Transformer-based models. It underscores the necessity of equipping fully autonomous vehicles with diverse sensors to ensure robust and reliable operation. The survey explores the advantages and drawbacks of cameras, LiDAR, and radar sensors. Additionally, it summarizes autonomy datasets and examines the latest advancements in multimodal fusion-based methods. The survey concludes by highlighting the ongoing challenges, open issues, and potential directions for future research.}, journal={World Electric Vehicle Journal}, author={Alaba, Simegnew Yihunie and Gurbuz, Ali Cafer and Ball, John}, year={2024}, month={Jan} }
@article{biswas_alam_gurbuz_2024, title={HRSpecNET: A Deep Learning-Based High-Resolution Radar Micro-Doppler Signature Reconstruction for Improved HAR Classification}, url={https://doi.org/10.1109/TRS.2024.3396172}, DOI={10.1109/TRS.2024.3396172}, abstractNote={Micro-Doppler signatures (μ-DS) are widely used for human activity recognition (HAR) using radar. However, traditional methods for generating μ-DS, such as the Short-Time Fourier Transform (STFT), suffer from limitations, such as the trade-off between time and frequency resolution, noise sensitivity, and parameter calibration. To address these limitations, we propose a novel deep learning-based approach to reconstruct high-resolution μ-DS directly from 1D complex time-domain signal. Our deep learning architecture consists of an autoencoder block to improve signal-to-noise ratio (SNR), an STFT block to learn frequency transformations to generate pseudo spectrograms, and finally, a UNET block to reconstruct high-resolution spectrogram images. We evaluated our proposed architecture on both synthetic and real-world data. For synthetic data, we generated 1D complex time domain signals with multiple time-varying frequencies and evaluated and compared the ability of our network to generate high-resolution μ-DS and perform in different SNR levels. For real-world data, a challenging radar-based American Sign Language (ASL) dataset consisting of 100 words was used to evaluate the classification performance achieved using the μ-DS generated by the proposed approach. The results showed that the proposed approach outperforms the classification accuracy of traditional STFT-based μ-DS by 3.48%. Both synthetic and experimental μ-DS show that the proposed approach learns to reconstruct higher-resolution and sparser spectrograms.}, journal={IEEE Transactions on Radar Systems}, author={Biswas, Sabyasachi and Alam, Ahmed Manavi and Gurbuz, Ali C.}, year={2024} }
@article{biswas_ayna_gurbuz_2024, title={PLFNets: Interpretable Complex-Valued Parameterized Learnable Filters for Computationally Efficient RF Classification}, url={https://doi.org/10.1109/TRS.2024.3486183}, DOI={10.1109/TRS.2024.3486183}, journal={IEEE Transactions on Radar Systems}, author={Biswas, Sabyasachi and Ayna, Cemre Omer and Gurbuz, Ali Cafer}, year={2024} }
@article{farhad_alam_biswas_rafi_gurbuz_kurum_2024, title={SDR-Based Dual Polarized L-Band Microwave Radiometer Operating From Small UAS Platforms}, url={https://doi.org/10.1109/JSTARS.2024.3394054}, DOI={10.1109/JSTARS.2024.3394054}, abstractNote={Passive microwave remote sensing is a vital tool for acquiring valuable information regarding the Earth's surface, with significant applications in agriculture, water management, forestry, and various environmental disciplines. Precision agricultural (PA) practices necessitate the availability of field-scale, high-resolution remote sensing data products. This study focuses on the design and development of a cost-effective, portable L-band microwave radiometer capable of operating from an unmanned aircraft system (UAS) platform to measure high-resolution surface brightness temperature ( $T_{B}$ ). This radiometer consists of a dual-polarized (Horizontal polarized, H-pol and Vertical polarized, V-pol) antenna and a software-defined radio (SDR)-based receiver system with a 30 MHz sampling rate. The post-processing methodology encompasses the conversion of raw in-phase and quadratic (I&Q) surface emissions to radiation $T_{B}$ through internal and external calibrations. Radiometric measurements were conducted over an experimental site covering both bare soil within an agricultural field and a large water body. The results yielded a high-resolution $T_{B}$ map that effectively delineated the boundaries between land and water, and identified land surface features. The radiometric temperature measurements of the sky and blackbody demonstrated a standard deviation of 0.95 K for H-pol and 0.57 K for V-pol in the case of the sky and 0.39 K for both H-pol and V-pol in the case of the blackbody observations. The utilization of (I&Q) samples acquired via the radiometer digital back-end facilitates the generation of different time-frequency (TF) analyses through short-time Fourier transform (STFT) and power spectral density (PSD). The transformation of radiometer samples into TF representations aids in the identification and mitigation of radio frequency interference (RFI) originating from the instrument itself and external sources.}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, author={Farhad, Md Mehedi and Alam, Ahmed Manavi and Biswas, Sabyasachi and Rafi, Mohammad Abdus Shahid and Gurbuz, Ali C. and Kurum, Mehmet}, year={2024} }
@article{farhad_kurum_gurbuz_2023, title={A Ubiquitous GNSS-R Methodology to Estimate Surface Reflectivity Using Spinning Smartphone Onboard a Small UAS}, url={https://doi.org/10.1109/JSTARS.2023.3294833}, DOI={10.1109/JSTARS.2023.3294833}, abstractNote={Global Navigation Satellite Systems (GNSS) Reflectometry (GNSS-R) has gained significant attention in retrieving geophysical parameters of the Earth's surface using ground, airborne, and spaceborne systems in the past decade. Such studies have mainly been investigated through custom-built systems or networks of geodetic receivers and antennas. For the broader adaptation of such an approach in precision agriculture or small-scale experiments, we have recently conjectured that a smartphone's built-in GNSS chip and antenna mounted on a small Unmanned Aircraft System (UAS) platform could be used to estimate the reflectivity of the soil surface using reflected GNSS signals. The main barrier to using a smartphone as a ubiquitous GNSS-R receiver is the built-in antenna's irregular radiation pattern that makes the measurement signal highly angular dependent. This study provides a unique and practical solution to lessen the impact of antenna radiation patterns on reflectivity estimation by spinning two smartphones mounted on two separate ground plate and taking the logarithmic difference of such simultaneous measurements. In this proposed configuration, a down-facing spinning smartphone on a UAS platform collects reflected signals. At the same time, another identical spinning smartphone is located on the ground, providing reference data in an open area. We recently conducted several experiments to test this concept, including various configurations on the ground and UAS platform in late 2021. First, we investigate the impact of the rotation speed on the received signal strength for roughly 4 hours. Second, we compare the results from measurements with the spinning smartphone on a small UAS and the ground. We also discuss the trade-offs involved in rotation and flight dynamics. Our findings show that a ubiquitous GNSS-R system based on spinning smartphones operating from a small UAS platform can estimate surface reflectivity at the sub-field scale.}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, author={Farhad, Md Mehedi and Kurum, Mehmet and Gurbuz, Ali Cafer}, year={2023} }
@article{biswas_ayna_gurbuz_gurbuz_2023, title={CV-SincNet: Learning Complex Sinc Filters From Raw Radar Data for Computationally Efficient Human Motion Recognition}, url={https://doi.org/10.1109/TRS.2023.3310894}, DOI={10.1109/TRS.2023.3310894}, abstractNote={The utilization of radio-frequency (RF) sensing in cyber-physical human systems, such as human-computer interfaces or smart environments, is an emerging application that requires real-time human motion recognition. However, current state-of-the-art radar-based recognition techniques rely on computing various RF data representations, such as range-Doppler or range-Angle maps, micro-Doppler signatures, or higher dimensional representations, which have great computational complexity. Consequently, classification of raw radar data has garnered increasing interest, while remaining limited in the accuracy that can be attained for recognition of even simple gross motor activities. To help address this challenge, this paper proposes a more interpretable complex-valued neural network design. Complex sinc filters are designed to learn frequency-based relationships directly from the complex raw radar data in the initial layer of the proposed model. The complex-valued sinc layer consists of windowed band-pass filters that learn the center frequency and bandwidth of each filter. A challenging RF dataset consisting of 100 words from American Sign Language (ASL) is selected to verify the model. About 40% improvement in classification accuracy was achieved over the application of a 1D CNN on raw RF data, while 8% improvement was achieved compared to real-valued SincNet. Our proposed approach achieved a 4% improvement in accuracy over that attained with a 2D CNN applied to micro-Doppler spectrograms, while also reducing the overall computational latency by 71%.}, journal={IEEE Transactions on Radar Systems}, author={Biswas, Sabyasachi and Ayna, Cemre Omer and Gurbuz, Sevgi Z. and Gurbuz, Ali C.}, year={2023} }
@article{ayna_mdrafi_du_gurbuz_2023, title={Learning-Based Optimization of Hyperspectral Band Selection for Classification}, url={https://www.mdpi.com/2072-4292/15/18/4460}, DOI={10.3390/rs15184460}, abstractNote={Hyperspectral sensors acquire spectral responses from objects with a large number of narrow spectral bands. The large volume of data may be costly in terms of storage and computational requirements. In addition, hyperspectral data are often information-wise redundant. Band selection intends to overcome these limitations by selecting a small subset of spectral bands that provide more information or better performance for particular tasks. However, existing band selection techniques do not directly maximize the task-specific performance, but rather utilize hand-crafted metrics as a proxy to the final goal of performance improvement. In this paper, we propose a deep learning (DL) architecture composed of a constrained measurement learning network for band selection, followed by a classification network. The proposed joint DL architecture is trained in a data-driven manner to optimize the classification loss along band selection. In this way, the proposed network directly learns to select bands that enhance the classification performance. Our evaluation results with Indian Pines (IP) and the University of Pavia (UP) datasets show that the proposed constrained measurement learning-based band selection approach provides higher classification accuracy compared to the state-of-the-art supervised band selection methods for the same number of bands selected. The proposed method shows 89.08% and 97.78% overall accuracy scores for IP and UP respectively, being 1.34% and 2.19% higher than the second-best method.}, journal={Remote Sensing}, author={Ayna, Cemre Omer and Mdrafi, Robiulhossain and Du, Qian and Gurbuz, Ali Cafer}, year={2023}, month={Sep} }
@article{alam_kurum_ogut_gurbuz_2024, title={Microwave Radiometer Calibration Using Deep Learning With Reduced Reference Information and 2-D Spectral Features}, url={https://doi.org/10.1109/JSTARS.2023.3333268}, DOI={10.1109/JSTARS.2023.3333268}, abstractNote={The accuracy of geophysical retrievals from radiometers relies on the quality of calibrations, which encompasses both absolute radiometric accuracy and spectral consistency. Radiometers have employed various calibration techniques, which include the utilization of external calibration targets, vicarious sources, and internal calibrators like noise diodes or matched reference loads. Calibration techniques pose several significant challenges such as frequency dependence, instrumental effects, environmental influences, drift, aging, and radio frequency interference. Recent advancements in hardware and processing units have enabled passive radiometers to collect raw samples of the observed scene that contain both temporal and spectral information. Leveraging advanced modeling techniques such as deep learning (DL) architecture can detect subtle correlations, non-linear dependencies, and higher-order interactions within the data. This capability allows them to extract valuable information that may have been difficult to capture using conventional methods. This study will utilize NASA's Soil Moisture Active Passive (SMAP) satellite's level 1A and level 1B data products to develop a DL-based radiometer calibrator to estimate antenna temperature. Spectrograms of second raw moments equivalent to power carrying the two-dimensional spectral features will be the primary input in a supervised convolutional neural network-based architecture. DL-based calibrator has demonstrated high R 2 and low root mean square error when incorporating spectral information from both reference and noise diodes and when not considering this information. The findings from this analysis will suggest that the ancillary features in DL-based calibrators such as internal thermistor temperature and loss elements exhibit sufficient accuracy in estimating antenna temperature to compensate for variations in receiver noise temperature and short-term gain fluctuations in the absence of the reference load and noise diode power. The proposed calibration technique with reduced reference information might enable radiometers for a higher number of antenna scene observations within a footprint.}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, author={Alam, Ahmed Manavi and Kurum, Mehmet and Ogut, Mehmet and Gurbuz, Ali C.}, year={2024} }
@article{nabi_senyurek_lei_kurum_gurbuz_2023, title={Quasi-Global Assessment of Deep Learning-Based CYGNSS Soil Moisture Retrieval}, url={https://doi.org/10.1109/JSTARS.2023.3287591}, DOI={10.1109/JSTARS.2023.3287591}, abstractNote={A high spatial and temporal resolution global soil moisture product is essential for understanding hydrologic and meteorological processes and enhancing agricultural applications. Global Navigation Satellite System (GNSS) signals at L-band frequencies that reflect off the land surface can convey high-resolution land surface information, including surface soil moisture (SM). Cyclone Global Navigation Satellite System (CYGNSS) constellation generates Delay-Doppler Maps (DDMs) that contain important Earth surface information from GNSS reflection measurements. DDMs are affected by soil moisture and other factors such as complex topography, soil texture, and overlying vegetation. Including entire DDM information can help reduce the uncertainty of SM estimation under different conditions along with remotely sensed geophysical data. This work extends our previously developed deep learning (DL) framework to a global scale by utilizing processed DDM measurements (analog power, effective scattering area, and bistatic radar cross-section) and ancillary data (elevation, slope, water percentage, soil properties and vegetation water content). The DL model is trained and evaluated using the Soil Moisture Active Passive (SMAP) mission's enhanced SM products at 9-km resolution. This study comprehensively evaluates the DL model against publicly available CYGNSS-based SM products at a quasi-global scale. In addition to the typical comparison against in-situ measurements, a robust triple collocation technique is used to evaluate the DL-based SM product and other CYGNSS-derived SM products.}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, author={Nabi, M M and Senyurek, Volkan and Lei, Fangni and Kurum, Mehmet and Gurbuz, Ali Cafer}, year={2023} }
@article{nabi_senyurek_gurbuz_kurum_2022, title={Deep Learning-Based Soil Moisture Retrieval in CONUS Using CYGNSS Delay–Doppler Maps}, url={https://doi.org/10.1109/JSTARS.2022.3196658}, DOI={10.1109/JSTARS.2022.3196658}, abstractNote={NASA's Cyclone Global Navigation Satellite System (CYGNSS) mission has gained significant attention within the land remote sensing community for estimating soil moisture (SM) by using the Global Navigation System Reflectometry (GNSS-R) technique. CYGNSS constellation generates Delay-Doppler Maps (DDM)s, containing important earth surface information from GNSS reflection measurements. Many previous studies considered only designed features from CYGNSS DDM such as the peak value of DDM, whereas the whole DDM image is affected by SM, topography, inundation, and overlying vegetation. In this paper, a deep learning (DL)-based framework is presented for estimating SM products in the Continental United States (CONUS) by leveraging spaceborne GNSS-R DDM observations provided by the CYGNSS constellation along with other remotely sensed geophysical data products. A data-driven approach utilizing convolutional neural networks (CNNs) is developed to determine complex relationships between the reflected measurements and surface parameters which can help to provide improved SM estimation. The CNN model is trained jointly with three types of processed DDM images of Analog Power, Effective scattering area, and Bistatic Radar Cross-section (BRCS) with other auxiliary geophysical information such as elevation, soil properties, normalized difference vegetation index (NDVI) and vegetation water content (VWC). The model is trained and evaluated using from the Soil Moisture Active Passive (SMAP) mission's enhanced SM products at a 9 km × 9 km resolution with VWC less than 5 kg/m$^{2}$. The mean unbiased root-mean-square difference (ubRMSD) between concurrent CYGNSS and SMAP SM retrievals from 2017 to 2020 is 0.0366 $m^{3}/m^{3}$ with a correlation coefficient of 0.93 over 5-fold cross-validation and 0.0333 $m^{3}/m^{3}$ with a correlation coefficient of 0.94 over year-based cross-validation at spatial resolution of 9 km × 9 km and temporal resolution same as the CYGNSS mission.}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, author={Nabi, M M and Senyurek, Volkan and Gurbuz, Ali C. and Kurum, Mehmet}, year={2022} }
@article{rahman_malaia_gurbuz_griffin_crawford_gurbuz_2022, title={Effect of Kinematics and Fluency in Adversarial Synthetic Data Generation for ASL Recognition With RF Sensors}, volume={58}, url={https://doi.org/10.1109/TAES.2021.3139848}, DOI={10.1109/TAES.2021.3139848}, abstractNote={RF sensors have been recently proposed as a new modality for sign language processing technology. They are noncontact, effective in the dark, and acquire a direct measurement of signing kinematic via exploitation of the micro-Doppler effect. First, this work provides an in depth comparative examination of the kinematic properties of signing as measured by RF sensors for both fluent ASL users and hearing imitation signers. Second, as ASL recognition techniques utilizing deep learning requires a large amount of training data, this work examines the effect of signing kinematics and subject fluency on adversarial learning techniques for data synthesis. The following two different approaches for the synthetic training data generation are proposed: 1) adversarial domain adaptation to minimize the differences between imitation signing and fluent signing data and 2) kinematically-constrained generative adversarial networks for accurate synthesis of RF signing signatures. The results show that the kinematic discrepancies between imitation signing and fluent signing are so significant that training on data directly synthesized from fluent RF signers offers greater performance (93% top-5 accuracy) than that produced by adaptation of imitation signing (88% top-5 accuracy) when classifying 100 ASL signs.}, number={4}, journal={IEEE Transactions on Aerospace and Electronic Systems}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Rahman, Mohammad Mahbubur and Malaia, Evie A. and Gurbuz, Ali Cafer and Griffin, Darrin J. and Crawford, Chris and Gurbuz, Sevgi Zubeyde}, year={2022}, month={Aug}, pages={2732–2745} }
@article{senyurek_farhad_gurbuz_kurum_adeli_2022, title={Fusion of Reflected GPS Signals With Multispectral Imagery to Estimate Soil Moisture at Subfield Scale From Small UAS Platforms}, url={https://doi.org/10.1109/JSTARS.2022.3197794}, DOI={10.1109/JSTARS.2022.3197794}, abstractNote={This study proposes a low-cost and “proof-of-concept” methodology to obtain high spatial resolution soil moisture (SM) via processing reflected Global Positioning System (GPS) and a multispectral camera data acquired by small Unmanned Aircraft System (UAS) platforms. An SM estimation model is developed using a random forest (RF) machine-learning (ML) algorithm by combining features obtained from reflected GPS signals (collected by smartphones and commercial off the shelf receivers) in conjunction with ancillary vegetation indices from the multispectral camera data. The proposed ML algorithm uses in-situ SM measurements acquired via SM probes as labels. A preliminary field experiment was conducted on 210 m by 110 m (2.31 ha) crop fields (corn and cotton) in 2020 (from January to November, including crop planting through senescence time period) at Mississippi State University (MSU)'s the heavily instrumented North Farm to acquire data needed for the ML model to train and test. Our results showed that both fields could be covered by GPS reflectometry measurements with about 13 minutes of flight time at a 15-m altitude, and SM can be mapped with 5m × 5m spatial resolution (corresponding to the elongated first Fresnel zone). The model is trained with and validated against eight in-situ SM station datasets via 10-fold and leave-one-probe-out cross-validation techniques. Overall root-mean-square errors (RMSE) of 0.013 m $^{3}$ m$^{-3}$ volumetric SM and R-value of 0.95 [-] are obtained for 10-fold cross-validation. The proposed model reached an RMSE of 0.033 m $^{3}$ m$^{-3}$ and an R-value of 0.5 [-] in leave-one-probe-out cross-validation. While having limited data, the results indicate that high resolution SM measurement can be achieved with a low-cost GPS reflectometry system onboard a small UAS platform for use in precision agriculture (PA) applications.}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, author={Senyurek, Volkan and Farhad, Md Mehedi and Gurbuz, Ali C. and Kurum, Mehmet and Adeli, Ardeshir}, year={2022} }
@article{alam_kurum_gurbuz_2022, title={Radio Frequency Interference Detection for SMAP Radiometer Using Convolutional Neural Networks}, url={https://doi.org/10.1109/JSTARS.2022.3223198}, DOI={10.1109/JSTARS.2022.3223198}, abstractNote={Passive remote sensing is a crucial technology for climate studies and Earth science. NASA's soil moisture active passive (SMAP) is a remote sensing observatory that uses passive microwave radiometer measurements to estimate soil moisture and detect the freeze or thaw state. Despite operating in the protected band of the radio spectrum (1400-1427 MHz), the radiometer's measurements are nonetheless tainted by radio frequency interference (RFI). An increasing number of radio-frequency transmissions such as those from air surveillance radars, 5G wireless communications, and unmanned aerial vehicles are contributing to RFI through either out-of-band emissions or operating in-band illegally. Physical modeling to detect RFI globally might prove to be challenging as RFI can be generated from single as well as multiple sources and these can be divided as pulsed or continuous wave RFI. In this study, a deep learning (DL) based RFI detection method is proposed with a novel convolutional neural network framework that can detect different types of RFI on a global scale. This is a data-driven approach where the detection framework learns directly from the SMAP data products to make a decision whether a certain footprint is RFI contaminated or not. SMAP's level 1A data products containing antenna counts of different raw moments along with Stokes parameters are used in this study to produce spectrograms and level 1B data products containing the quality flags are used to dynamically label those spectrograms. This study's robust DL framework provided the highest accuracy with the raw moments of horizontal polarization (99.99%) to detect RFI globally.}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, author={Alam, Ahmed Manavi and Kurum, Mehmet and Gurbuz, Ali C.}, year={2022} }
@article{kurtoglu_gurbuz_malaia_griffin_crawford_gurbuz_2022, title={ASL Trigger Recognition in Mixed Activity/Signing Sequences for RF Sensor-Based User Interfaces}, volume={52}, url={https://doi.org/10.1109/THMS.2021.3131675}, DOI={10.1109/THMS.2021.3131675}, abstractNote={The past decade has seen great advancements in speech recognition for control of interactive devices, personal assistants, and computer interfaces. However, deaf and hard-of-hearing (HoH) individuals, whose primary mode of communication is sign language, cannot use voice-controlled interfaces. Although there has been significant work in video-based sign language recognition, video is not effective in the dark and has raised privacy concerns in the deaf community when used in the context of human ambient intelligence. RF sensors have been recently proposed as a new modality that can be effective under the circumstances where video is not. This article considers the problem of recognizing a trigger sign (wake word) in the context of daily living, where gross motor activities are interwoven with signing sequences. The proposed approach exploits multiple RF data domain representations (time-frequency, range-Doppler, and range-angle) for sequential classification of mixed motion data streams. The recognition accuracy of signs with varying kinematic properties is compared and used to make recommendations on appropriate trigger sign selection for RF-sensor-based user interfaces. The proposed approach achieves a trigger sign detection rate of 98.9% and a classification accuracy of 92% for 15 ASL words and three gross motor activities.}, number={4}, journal={IEEE Transactions on Human-Machine Systems}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Kurtoglu, Emre and Gurbuz, Ali C. and Malaia, Evie A. and Griffin, Darrin and Crawford, Chris and Gurbuz, Sevgi Z.}, year={2022}, month={Aug}, pages={699–712} }
@article{senyurek_gurbuz_kurum_2021, title={Assessment of Interpolation Errors of CYGNSS Soil Moisture Estimations}, volume={14}, url={https://doi.org/10.1109/JSTARS.2021.3113565}, DOI={10.1109/JSTARS.2021.3113565}, abstractNote={High spatio-temporal soil moisture (SM) is essential for many meteorological, hydrological, and agricultural applications and studies. Spaceborne Global Navigation Satellite System Reflectometry (GNSS-R) provides a promising opportunity for high-resolution SM retrievals. NASAs Cyclone Global Navigation Satellite System (CYGNSS) is a recent GNSS-R application that offers relatively high spatial and temporal resolution observations from Earths surface. However, the quasi-random sampling of land surface by the CYGNSS constellation circumvents obtaining fully observed daily SM predictions at high spatial resolutions. Spatial interpolation techniques may fill this gap and provide a fully covered high-resolution daily SM estimation. However, the spatial interpolation errors need to be assessed when applied to the quasi-random 9-km CYGNSS based SM estimations. In this paper, we conduct interpolation error analysis using the SMAP Enhanced L3 Radiometer Global Daily 9-km product, sampled at the CYGNSS observation locations. The results indicate that the overall interpolation error (RMSE) was 0.013 m3m-3 over SMAPs recommended grids. In addition, sparse CYGNSS SM observations are directly interpolated. The achieved results show that interpolated and observed CYGNSS SM values have similar performance metrics when validated with the SMAP 9- km gridded SM product as well as sparse soil moisture networks.}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Senyurek, Volkan and Gurbuz, Ali Cafer and Kurum, Mehmet}, year={2021}, pages={9815–9825} }
@article{gurbuz_rahman_kurtoglu_malaia_gurbuz_griffin_crawford_2022, title={Multi-Frequency RF Sensor Fusion for Word-Level Fluent ASL Recognition}, url={https://doi.org/10.1109/JSEN.2021.3078339}, DOI={10.1109/JSEN.2021.3078339}, abstractNote={Deaf spaces are unique indoor environments designed to optimize visual communication and Deaf cultural expression. However, much of the technological research geared towards the deaf involve use of video or wearables for American sign language (ASL) translation, with little consideration for Deaf perspective on privacy and usability of the technology. In contrast to video, RF sensors offer the avenue for ambient ASL recognition while also preserving privacy for Deaf signers. Methods: This paper investigates the RF transmit waveform parameters required for effective measurement of ASL signs and their effect on word-level classification accuracy attained with transfer learning and convolutional autoencoders (CAE). A multi-frequency fusion network is proposed to exploit data from all sensors in an RF sensor network and improve the recognition accuracy of fluent ASL signing. Results: For fluent signers, CAEs yield a 20-sign classification accuracy of %76 at 77 GHz and %73 at 24 GHz, while at X-band (10 Ghz) accuracy drops to 67%. For hearing imitation signers, signs are more separable, resulting in a 96% accuracy with CAEs. Further, fluent ASL recognition accuracy is significantly increased with use of the multi-frequency fusion network, which boosts the 20-sign fluent ASL recognition accuracy to 95%, surpassing conventional feature level fusion by 12%. Implications: Signing involves finer spatiotemporal dynamics than typical hand gestures, and thus requires interrogation with a transmit waveform that has a rapid succession of pulses and high bandwidth. Millimeter wave RF frequencies also yield greater accuracy due to the increased Doppler spread of the radar backscatter. Comparative analysis of articulation dynamics also shows that imitation signing is not representative of fluent signing, and not effective in pre-training networks for fluent ASL classification. Deep neural networks employing multi-frequency fusion capture both shared, as well as sensor-specific features and thus offer significant performance gains in comparison to using a single sensor or feature-level fusion.}, journal={IEEE Sensors Journal}, author={Gurbuz, Sevgi Z. and Rahman, M. Mahbubur and Kurtoglu, Emre and Malaia, Evie and Gurbuz, Ali Cafer and Griffin, Darrin J. and Crawford, Chris}, year={2022}, month={Jun} }
@article{mdrafi_du_gurbuz_tang_ma_younan_2020, title={Attention-Based Domain Adaptation Using Residual Network for Hyperspectral Image Classification}, volume={13}, url={https://doi.org/10.1109/JSTARS.2020.3035382}, DOI={10.1109/JSTARS.2020.3035382}, abstractNote={In remote sensing images, domain adaptation (DA) deals with the regions where labeling information is unknown. Typically, hand-driven features for learning a common distribution among known and unknown regions have been extensively exploited to perform the classification task in hyperspectral images with the aid of state-of-the-art machine learning algorithms. Under limited training samples and using hand-crafted features, the classification performance degrades significantly. To overcome the engineered feature extraction process, an automatic feature extraction scheme can be seen useful to generate more complex but useful features for classification. Deep-learning-based architectures have been found to be pivotal on this regard. Deep learning algorithms are effectively used in hyperspectral domain to solve the DA problem. However, attention-based activation mappings, which are very successful for distinguishing different classes of images via transferring relevant mappings from a deep-to-shallow network is not widely explored in DA domain. In this article, we have opted to use attention-based DA through transferring different levels of attentions by means of different types of activation mappings from a deep residual teacher network to a shallow residual student network. Our goal is to provide useful but more complex features to the shallow student network for improving the overall classification in case of DA task. It has been shown that for different kinds of activation mappings, the proposed attention-based transfer improves the performance of the shallow network for the DA problem. It also outperforms the state-of-the-art DA methods based on traditional machine learning and deep learning paradigms.}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Mdrafi, Robiulhossain and Du, Qian and Gurbuz, Ali Cafer and Tang, Bo and Ma, Li and Younan, Nicolas H.}, year={2020}, pages={6424–6433} }
@article{boyd_gurbuz_kurum_garrison_nold_piepmeier_vega_bindlish_2020, title={Cramer–Rao Lower Bound for SoOp-R-Based Root-Zone Soil Moisture Remote Sensing}, volume={13}, url={https://doi.org/10.1109/JSTARS.2020.3029158}, DOI={10.1109/JSTARS.2020.3029158}, abstractNote={Signals of opportunity (SoOp) reflectometry (SoOp-R) is a maturing field for geophysical remote sensing as evidenced by the growing number of airborne and spaceborne experiments. As this approach receives more attention, it is worth analyzing SoOp-R's capabilities to retrieve subsurface soil moisture (SM) by leveraging communication and navigation satellite transmitters. In this research, the Cramer-Rao lower bound (CRLB) is used to identify the effects of variable SoOp-R parameters on the best achievable estimation error for root-zone soil moisture (RZSM). This study investigates the use of multiple frequency, polarization, and incidence angle measurement configurations on a two-layered dielectric profile. The results also detail the effects of variable SM conditions on the capability of SoOp-R systems to predict subsurface SM. The most prevalent observation is the importance of using at least two frequencies to limit uncertainties from subsurface SM estimates. If at least two frequencies are used, the CRLB of a profile is retrievable within the root-zone depending on the surface SM content as well as the number of independent measurements of the profile. For a depth of 30 cm, it is observed that a CRLB corresponding to 4% RZSM estimation accuracy is achievable with as few as two dual-frequency-based SoOp-R measurements. For this depth, increasing number of measurements provided by polarization and incidence angle allow for sensing of increasingly wet SM profile structures. This study, overall, details a methodology by which SoOp-R receiver system can be designed to achieve a desired CRLB using a tradeoff study between the available measurements and SM profile.}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Boyd, Dylan Ray and Gurbuz, Ali Cafer and Kurum, Mehmet and Garrison, James L. and Nold, Benjamin R. and Piepmeier, Jeffrey R. and Vega, Manuel and Bindlish, Rajat}, year={2020}, pages={6101–6114} }
@article{senyurek_lei_boyd_gurbuz_kurum_moorhead_2020, title={Evaluations of Machine Learning-Based CYGNSS Soil Moisture Estimates against SMAP Observations}, volume={12}, url={https://doi.org/10.3390/rs12213503}, DOI={10.3390/rs12213503}, abstractNote={This paper presents a machine learning (ML) framework to derive a quasi-global soil moisture (SM) product by direct use of the Cyclone Global Navigation Satellite System (CYGNSS)’s high spatio-temporal resolution observations over the tropics (within ±38° latitudes) at L-band. The learning model is trained by using in-situ SM data from the International Soil Moisture Network (ISMN) sites and various space-borne ancillary data. The approach produces daily SM retrievals that are gridded to 3 km and 9 km within the CYGNSS spatial coverage. The performance of the model is independently evaluated at various temporal scales (daily, 3-day, weekly, and monthly) against Soil Moisture Active Passive (SMAP) mission’s enhanced SM products at a resolution of 9 km × 9 km. The mean unbiased root-mean-square difference (ubRMSD) between concurrent (same calendar day) CYGNSS and SMAP SM retrievals for about three years (from 2017 to 2019) is 0.044 cm3 cm−3 with a correlation coefficient of 0.66 over SMAP recommended grids. The performance gradually improves with temporal averaging and degrades over regions regularly flagged by SMAP such as dense forest, high topography, and coastlines. Furthermore, CYGNSS and SMAP retrievals are evaluated against 170 ISMN in-situ observations that result in mean unbiased root-mean-square errors (ubRMSE) of 0.055 cm3 cm−3 and 0.054 cm3 cm−3, respectively, and a higher correlation coefficient with CYGNSS retrievals. It is important to note that the proposed approach is trained over limited in-situ observations and is independent of SMAP observations in its training. The retrieval performance indicates current applicability and future growth potential of GNSS-R-based, directly measured spaceborne SM products that can provide improved spatio-temporal resolution than currently available datasets.}, number={21}, journal={Remote Sensing}, publisher={MDPI AG}, author={Senyurek, Volkan and Lei, Fangni and Boyd, Dylan and Gurbuz, Ali Cafer and Kurum, Mehmet and Moorhead, Robert}, year={2020}, month={Oct}, pages={3503} }
@article{kurum_farhad_gurbuz_2021, title={Integration of Smartphones Into Small Unmanned Aircraft Systems to Sense Water in Soil by Using Reflected GPS Signals}, volume={14}, url={https://doi.org/10.1109/JSTARS.2020.3041162}, DOI={10.1109/JSTARS.2020.3041162}, abstractNote={We investigate the feasibility of using built-in GNSS sensors within ubiquitous smartphone devices from a small UAS for the purpose of land remote sensing. We summarize the experimental findings and challenges that need to be resolved in order to perform the GNSS reflectometry (GNSS-R) technique via smartphones. In late 2018, a series of experiments were conducted and designed by integrating two smartphones into a multicopter UAS by attaching them to ground plates to isolate and record both direct and reflected GNSS carrier-to-noise density ratio ($C/N_0$) separately. It was demonstrated that, first, fluctuations of moving GNSS specular reflections are correlated with spatial ground features with appreciable dynamic range and second, radiation pattern of the smartphone's inbuilt antenna has a significant effect on the received signal strength. In 2020, more experiments were conducted to examine the quality of in-built chip and antenna of a smartphone with regard to the GNSS-R approach as well as the consistency of measurements. These follow-up experiments involved, first, placement of the smartphone on a pan-tilt mechanism on a tripod, second, formation flights with smartphone on a gimbal and a high-quality custom-built dual-channel GNSS-R receiver, and, third, flying the UAS at different times of the day on two consecutive days. It was demonstrated that, first, the radiation pattern of the smartphone's GNSS antenna are observed to be highly irregular, but time-invariant, and, second, internal GNSS chip produces observables of sufficient quality, and, third, the fluctuations of the reflected signal are repeatable under the same configuration at different times.}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Kurum, Mehmet and Farhad, Md. Mehedi and Gurbuz, Ali Cafer}, year={2021}, pages={1048–1059} }
@article{mdrafi_gurbuz_2020, title={Joint Learning of Measurement Matrix and Signal Reconstruction via Deep Learning}, volume={6}, url={http://dx.doi.org/10.1109/tci.2020.2983153}, DOI={10.1109/tci.2020.2983153}, abstractNote={In this work, we propose an automatic sensing and reconstruction scheme based on deep learning within the compressive sensing (CS) framework. Classical CS utilizes pre-determined linear projections in the form of random measurements and convex optimization with a known sparsity basis to reconstruct signals. Here, we develop a data-driven approach to learn both the measurement matrix and the inverse reconstruction scheme for a given class of signals, such as images. The developed deep learning approach paves the way for end-to-end learning and reconstruction of signals with the aid of cascaded fully connected and multistage convolutional layers with a weighted loss function in an adversarial learning framework. Results obtained over the CIFAR-10 image database show that the proposed deep learning architectures provide higher peak signal-to-noise ratio (PSNR) levels, and, hence, learn better measurement matrices than that of randomly selected, specifically designed to reduce average coherence with a given basis, or state-of-the-art data driven approaches. The learned measurement matrices achieve higher PSNR compared to random or designed matrices not only when they are utilized in the proposed data-driven approach but also when used in $\ell _1$ based recovery. The reconstruction performance on the test dataset improves as more training samples are utilized. Quantitative results for sparsity level analysis, incremental measurement design, and various training scenarios are provided.}, journal={IEEE Transactions on Computational Imaging}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Mdrafi, Robiulhossain and Gurbuz, Ali Cafer}, year={2020}, pages={818–829} }
@article{senyurek_lei_boyd_kurum_gurbuz_moorhead_2020, title={Machine Learning-Based CYGNSS Soil Moisture Estimates over ISMN sites in CONUS}, volume={12}, url={https://doi.org/10.3390/rs12071168}, DOI={10.3390/rs12071168}, abstractNote={Soil moisture (SM) derived from satellite-based remote sensing measurements plays a vital role for understanding Earth’s land and near-surface atmosphere interactions. Bistatic Global Navigation Satellite System (GNSS) Reflectometry (GNSS-R) has emerged in recent years as a new domain of microwave remote sensing with great potential for SM retrievals, particularly at high spatio-temporal resolutions. In this work, a machine learning (ML)-based framework is presented for obtaining SM data products over the International Soil Moisture Network (ISMN) sites in the Continental United States (CONUS) by leveraging spaceborne GNSS-R observations provided by NASA’s Cyclone GNSS (CYGNSS) constellation alongside remotely sensed geophysical data products. Three widely-used ML approaches—artificial neural network (ANN), random forest (RF), and support vector machine (SVM)—are compared and analyzed for the SM retrieval through utilizing multiple validation strategies. Specifically, using a 5-fold cross-validation method, overall RMSE values of 0.052, 0.061, and 0.065 cm3/cm3 are achieved for the RF, ANN, and SVM techniques, respectively. In addition, both a site-independent and a year-based validation techniques demonstrate satisfactory accuracy of the proposed ML model, suggesting that this SM approach can be generalized in space and time domains. Moreover, the achieved accuracy can be further improved when the model is trained and tested over individual SM networks as opposed to combining all available SM networks. Additionally, factors including soil type and land cover are analyzed with respect to their impacts on the accuracy of SM retrievals. Overall, the results demonstrated here indicate that the proposed technique can confidently provide SM estimates over lightly-vegetated areas with vegetation water content (VWC) less than 5 kg/m2 and relatively low spatial heterogeneity.}, number={7}, journal={Remote Sensing}, publisher={MDPI AG}, author={Senyurek, Volkan and Lei, Fangni and Boyd, Dylan and Kurum, Mehmet and Gurbuz, Ali Cafer and Moorhead, Robert}, year={2020}, month={Apr}, pages={1168} }
@article{anjinappa_gurbuz_yapici_guvenc_2020, title={Off-Grid Aware Channel and Covariance Estimation in mmWave Networks}, volume={68}, url={https://doi.org/10.1109/TCOMM.2020.2980829}, DOI={10.1109/TCOMM.2020.2980829}, abstractNote={The spectrum scarcity at sub-6 GHz spectrum has made millimeter-wave (mmWave) frequency band a key component of the next-generation wireless networks. While mmWave spectrum offers extremely large transmission bandwidths to accommodate ever-increasing data rates, unique characteristics of this new spectrum need special consideration to achieve the promised network throughput. In this work, we consider the off-grid targets (basis mismatch) problem for mmWave communications. The off-grid effect naturally appears in compressed sensing (CS) techniques adopting a discretization approach for representing the angular domain. This approach yields a finite set of discrete angle points, which are an approximation to the continuous angular space, and hence degrade the accuracy of related parameter estimation. In order to cope with the off-grid effect, we present a novel parameter-perturbation framework to efficiently estimate the channel and the covariance for mmWave networks. The proposed algorithms employ a smart perturbation mechanism in conjunction with a low-complexity greedy framework of simultaneous orthogonal matching pursuit (SOMP), and jointly solve for the off-grid parameters and weights. Numerical results show a significant performance improvement through our novel framework as a result of handling the off-grid effects, which is totally ignored in the conventional sparse mmWave channel or covariance estimation algorithms.}, number={6}, journal={IEEE Transactions on Communications}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Anjinappa, Chethan Kumar and Gurbuz, Ali Cafer and Yapici, Yavuz and Guvenc, Ismail}, year={2020}, month={Jun}, pages={3908–3921} }
@article{boyd_kurum_eroglu_gurbuz_garrison_nold_vega_piepmeier_bindlish_2020, title={SCoBi Multilayer: A Signals of Opportunity Reflectometry Model for Multilayer Dielectric Reflections}, url={https://doi.org/10.3390/rs12213480}, DOI={10.3390/rs12213480}, abstractNote={A multilayer module is incorporated into the Signals of Opportunity (SoOp) Coherent Bistatic Scattering model (SCoBi) for determining the reflections and propagation of electric fields within a series of multilayer dielectric slabs. This module can be used in conjunction with other SCoBi components to simulate complex, bistatic simulation schemes that include features such as surface roughness, vegetation, antenna effects, and multilayer soil moisture interactions on reflected signals. This paper introduces the physics underlying the multilayer module and utilizes it to perform a simulation study of the response of SoOp-R measurements with respect to subsurface soil moisture parameters. For a frequency range of 100–2400 MHz, it is seen that the SoOp-R response to a single dielectric slab is mostly frequency insensitive; however, the SoOp-R response to multilayer dielectric slabs will vary between frequencies. The relationship between SoOp-R reflectivity and the contributing depth is visualized, and the results show that SoOp-R measurements can display sensitivity to soil moisture below the penetration depth. By simulation of simple soil moisture profiles with different wetting and drying gradients, the dielectric contrast between layers is shown to be the greatest contributing factor to subsurface soil moisture sensitivity. Overall, it is observed that different frequencies can sense different areas of a soil moisture profile, and this behavior can enable subsurface soil moisture data products from SoOp-R observations.}, journal={Remote Sensing}, author={Boyd, Dylan and Kurum, Mehmet and Eroglu, Orhan and Gurbuz, Ali Cafer and Garrison, James L. and Nold, Benjamin R. and Vega, Manuel A. and Piepmeier, Jeffrey R. and Bindlish, Rajat}, year={2020}, month={Oct} }
@article{ucer_kisacikoglu_yuksel_gurbuz_2019, title={An Internet-Inspired Proportional Fair EV Charging Control Method}, volume={13}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85068711622&partnerID=MN8TOARS}, DOI={10.1109/JSYST.2019.2903835}, abstractNote={Transportation systems are undergoing a major transition with the integration of electric vehicles (EVs). However, due to increase in battery energy and charger power ratings, potential adverse effects on the distribution grid is a crucial issue to be addressed. Large voltage drops at charging nodes will deteriorate the quality of power service and cause unfair utilization of grid capacity among EV users. Safe and efficient operation of the grid along with a fast, convenient, and fair charging strategy is an important research problem. In this paper, we adapt the additive increase multiplicative decrease (AIMD) algorithm used in the Internet congestion control to EV charging using only local node measurements. We analyze the relationship between distance and grid voltage, and show how to extract this information from local measurements. Then, we present a detailed analysis to understand the relationship between distance and charging power in a distribution network to better address the fairness in the proposed AIMD EV charging algorithm. Results show that localized information at charging node voltages include important signature information on grid congestion and can be used to implement AIMD control for EV charging.}, number={4}, journal={IEEE Systems Journal}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Ucer, Emin and Kisacikoglu, Mithat C. and Yuksel, Murat and Gurbuz, Ali Cafer}, year={2019}, pages={4292–4302} }
@article{gurbuz_cetiner_2020, title={CRLB based mode selection and enhanced DOA estimation for multifunctional reconfigurable arrays}, volume={38}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85074153583&partnerID=MN8TOARS}, DOI={10.1016/j.phycom.2019.100894}, journal={Physical Communication}, author={Gurbuz, A.C. and Cetiner, B.A.}, year={2020} }
@inproceedings{rafi_gurbuz_2019, title={Data Driven Measurement Matrix Learning for Sparse Reconstruction}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85069470178&partnerID=MN8TOARS}, DOI={10.1109/DSW.2019.8755557}, abstractNote={In this work, we learn jointly how to sense and reconstruct a class of signals such as images through a deep learning structure within the compressive sensing (CS) framework. We develop a data driven approach and learn both the measurement matrix and the inverse reconstruction scheme instead of utilizing random linear projections as measurements and reconstruction via convex optimization with a given known sparsity in conventional CS framework. To achieve this goal, we have designed an end to end deep neural network structure consisting of fully connected layers with cascaded convolutional layers to be trained and tested over a publicly available image dataset. Results show that the proposed technique provides higher peak signal to noise ratio (PSNR) levels and hence learns better measurement matrices than that of the randomly selected or specifically designed for a known sparsity basis to reduce average coherence. The reconstruction performance on the test dataset also gets better as more train samples are utilized. We also observe that the learned measurement matrices achieve higher PSNR compared to random or designed cases when they are used in ℓ 1 based recovery. Proposed reconstruction scheme has much less computational complexity compared to ℓ 1 minimization based reconstruction with comparable results.}, booktitle={2019 IEEE Data Science Workshop, DSW 2019 - Proceedings}, author={Rafi, R.H.M. and Gurbuz, A.C.}, year={2019}, pages={253–257} }
@inproceedings{rogers_ball_gurbuz_2019, title={Estimating the number of sources via deep learning}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85073122572&partnerID=MN8TOARS}, DOI={10.1109/RADAR.2019.8835819}, abstractNote={This paper presents a deep-learning-based approach to estimating the number of sources in a narrowband angle of arrival (AoA) estimation problem where all signals are coherent. This is an important problem in radar, sonar and communication systems, as many AoA estimators require accurate estimates of the number of sources. Herein, a 15-layer deep learning network with parametric rectified linear units and batch normalization layers is trained using the eigenvalues and the spatially smoothed covariance matrix entries as an input. Our initial results show that the proposed learning based system produces more accurate of estimates of the sources compared to the state-of-the-art techniques.}, booktitle={2019 IEEE Radar Conference, RadarConf 2019}, author={Rogers, J. and Ball, J.E. and Gurbuz, A.C.}, year={2019} }
@article{eroglu_kurum_boyd_gurbuz_2019, title={High spatio-temporal resolution cygnss soil moisture estimates using artificial neural networks}, volume={11}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85073447906&partnerID=MN8TOARS}, DOI={10.3390/rs11192272}, abstractNote={This paper presents a learning-based, physics-aware soil moisture (SM) retrieval algorithm for NASA’s Cyclone Global Navigation Satellite System (CYGNSS) mission. The goal of the proposed novel method is to advance CYGNSS-based SM estimations, exploiting the spatio-temporal resolution of the GNSS reflectometry (GNSS-R) signals to its highest potential within a machine learning framework. The methodology employs a fully connected Artificial Neural Network (ANN) regression model to perform SM predictions through learning the nonlinear relations of SM and other land geophysical parameters to the CYGNSS observables. In situ SM measurements from several International SM Network (ISMN) sites are used as reference labels; CYGNSS incidence angles, derived reflectivity and trailing edge slope (TES) values, as well as ancillary data, are exploited as input features for training and validation of the ANN model. In particular, the utilized ancillary data consist of normalized difference vegetation index (NDVI), vegetation water content (VWC), terrain elevation, terrain slope, and h-parameter (surface roughness). Land cover classification and inland water body masks are also used for the intermediate derivations and quality control purposes. The proposed algorithm assumes uniform SM over a 0.0833 ∘ × 0.0833 ∘ (approximately 9 km × 9 km around the equator) lat/lon grid for any CYGNSS observation that falls within this window. The proposed technique is capable of generating sub-daily and high-resolution SM predictions as it does not rely on time-series or spatial averaging of the CYGNSS observations. Once trained on the data from ISMN sites, the model is independent from other SM sources for retrieval. The estimation results obtained over unseen test data are promising: SM predictions with an unbiased root mean squared error of 0.0544 cm 3 /cm 3 and Pearson correlation coefficient of 0.9009 are reported for 2017 and 2018.}, number={19}, journal={Remote Sensing}, publisher={MDPI AG}, author={Eroglu, Orhan and Kurum, Mehmet and Boyd, Dylan and Gurbuz, Ali Cafer}, year={2019}, pages={2272} }
@inproceedings{rafi_gurbuz_2019, title={Learning to sense and reconstruct a class of signals}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85069445754&partnerID=MN8TOARS}, DOI={10.1109/RADAR.2019.8835706}, abstractNote={In this paper, we propose a deep learning structure to jointly learn both how to sense and reconstruct a class of signals. In contrast to classical compressive sensing (CS) framework that utilizes pre-determined linear projections as measurements and convex optimization with a known sparsity basis to reconstruct the signal, instead we develop a data driven approach and learn both the measurement matrix and the inverse reconstruction scheme. To achieve this, an end to end deep neural network with fully connected and convolutional layers are designed and trained over an image dataset. Our initial results show that the measurement matrix learned through the proposed technique provides higher peak signal to noise ratio (PSNR) levels compared to both randomly selected matrices or designed measurement matrices for an assumed sparsity basis for the dataset. Learned measurement matrices are tested in both $\ell_{1}$ minimization based sparse recovery and deep neural network structures and for both recovery schemes highest PSNR values are obtained with the learned measurement matrix. The $\ell_{1}$ based recovery achieves higher PSNR results compared to inversion with deep neural network, their results are comparable.}, booktitle={2019 IEEE Radar Conference, RadarConf 2019}, author={Rafi, R.H.M. and Gurbuz, A.C.}, year={2019} }
@inproceedings{gogineni_simpson_yan_o?neill_sood_gurbuz_gurbuz_2018, title={A CubeSat train for radar sounding and imaging of Antarctic ice sheet}, volume={2018-July}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85064247277&partnerID=MN8TOARS}, DOI={10.1109/IGARSS.2018.8519162}, abstractNote={In spite of more than 50 years of airborne radar soundings of Antarctic ice by the international community, there are still large gaps in ice thickness data. We propose a CubeSat satellite mission for complete sounding and imaging of Antarctica with 50 CubeSats integrated with a VHF radar system to sound the ice and image the ice-bed. One of the major challenges in orbital sounding of ice is off-vertical surface clutter that masks weak ice-bed echoes. We must obtain fine resolution both in the along track and cross track directions to reduce surface clutter. We can obtain fine resolution in the along track direction by synthesizing a large aperture by taking advantage of the forward motion of a satellite. However, we need a large antenna-array to obtain fine resolution in the cross track direction. We propose a train of 50 CubeSats with optimized offset position to obtain a 500-m long aperture and also coherently combine data from multiple passes of the train to obtain a very large aperture of 1–2 km in the cross track direction. Our initial analysis shows that we can obtain measurements with horizontal resolution of about 200 m and vertical resolution of about 20 m. The CubeSat will carry a transmitter and receiver with peak transmit power of about 50 W. We will synchronize all transmitters and receivers with a Ka-band system that serves as a communication link between the earth and Cubesats to downlink data and as command and control for the CubeSats.}, booktitle={International Geoscience and Remote Sensing Symposium (IGARSS)}, author={Gogineni, P. and Simpson, C.R. and Yan, J.-B. and O?Neill, C.R. and Sood, R. and Gurbuz, S.Z. and Gurbuz, A.C.}, year={2018}, pages={4138–4141} }
@inproceedings{gurbuz_2018, title={Adaptive measurement design for direction of arrival estimation and target tracking}, volume={10658}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85050849616&partnerID=MN8TOARS}, DOI={10.1117/12.2305048}, abstractNote={Compressive sensing theory states that a sparse vector x in dictionary A can be recovered from measurements y = WAx. For recovery of x, the measurement matrix W is generally chosen as random since a random W is sufficiently incoherent with a given basis A with high probability. Although Gaussian or Bernoulli random measurement matrices satisfy recovery requirements, they do not necessarily yield the best performance in terms of minimal mutual coherence or best parameter estimation. In literature several studies focused on measurement matrix design mainly to minimize some form of coherence between W and A to minimize measurement numbers while exact reconstruction is guaranteed. On the other hand, for enhanced parameter estimation W can be designed to minimize the Cramer Rao Lower Bound (CRLB). In this study, we propose direct and sequential measurement designs that minimizes the CRLB for the application of direction of arrival (DoA) estimation. Based on our results an adaptive target tracking procedure for single and multiple target scenarios is also proposed. Initial simulations show that measurement design solutions provide enhanced parameter estimation and target tracking performance compared to widely used random matrices in compressive sensing.}, booktitle={Proceedings of SPIE - The International Society for Optical Engineering}, author={Gurbuz, A.C.}, year={2018} }
@inproceedings{gurbuz_gürbüz_cetiner_2018, title={Cognitive radar utilizing multifunctional reconfigurable antennas}, volume={10633}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85050035183&partnerID=MN8TOARS}, DOI={10.1117/12.2304397}, abstractNote={Cognitive radar is a novel concept for next-generation radar systems, which as part of the perception-action cycle to improve the measurement process based on dynamic changes in the environment. Although most work in this area to-date have focused on adaptation on the transmitted waveform, in this paper, we propose adaptive control of novel multifunctional reconfigurable antennas (MRAs) as a mechanism for action within the cognitive radar framework. Reconfigurable parasitic layer based MRAs have the capability of dynamically and simultaneously changing its electromagnetic characteristics (mode of operation), e.g. antenna beam pattern, polarization, center frequency, or a combination of thereof. Different modes of an MRA are controlled via RF switches interconnecting the pixels of the reconfigurable parasitic layer. This enhanced capability can be controlled using adaptive mode selection schemes. In particular, an array of MRAs provides more degrees of freedom, where each element of an array can be controlled to generate one of many modes depending on the environmental measured variables as a feedback mechanism. In this work, a designed and fabricated reconfigurable parasitic layer based MRA operating over 4.94-4.99 GHz band with 25 different radiation patterns, i.e., modes of operation, is utilized for cognitive direction-of-arrival (DoA) estimation and target tracking. A novel computationally efficient iterative mode selection (IMS) technique for MRA arrays is developed, where the modes are cognitively selected to minimize the DoA estimation error in target track. It is demonstrated that the proposed cognitive mode selection for MRA arrays achieves remarkably lower estimation errors compared to uniform pattern arrays without adaptive capability.}, booktitle={Proceedings of SPIE - The International Society for Optical Engineering}, author={Gurbuz, A.C. and Gürbüz, S.Z. and Cetiner, B.}, year={2018} }
@inproceedings{gurbuz_cetiner_2018, title={Multifunctional reconfigurable antennas for cognitive radars}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85049937795&partnerID=MN8TOARS}, DOI={10.1109/RADAR.2018.8378790}, abstractNote={Cognitive radar systems offer improved performance through mimicking the perception-action cycle in human cognition. While most of the efforts in the area to-date focus on agile waveform design and sensor management, in this paper, we propose adaptive control of novel multifunctional reconfigurable antennas (MRAs) as a mechanism for action within the cognitive radar framework. Reconfigurable parasitic layer based MRAs have the capability of dynamically and simultaneously changing its electromagnetic characteristics (mode of operation), e.g. antenna beam pattern, polarization, center frequency, or a combination of thereof. In this work, a designed and fabricated reconfigurable parasitic layer based MRA is presented. A computationally efficient mode selection scheme is proposed to adaptively select the modes of the MRA for a defined cost such as improved SNR or lower Cramer Rao Bound. A cognitive target tracking framework is proposed through MRA mode selection for cognitive radars. Initial results show that MRAs with mode selection provides enhanced direction-of-arrival (DoA) estimation and cognitive target tracking performance.}, booktitle={2018 IEEE Radar Conference, RadarConf 2018}, author={Gurbuz, A.C. and Cetiner, B.}, year={2018}, pages={1510–1515} }
@inproceedings{gurbuz_2018, title={Perturbation based sparse subspace clustering}, volume={10658}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85050877754&partnerID=MN8TOARS}, DOI={10.1117/12.2304723}, abstractNote={Many application areas including signal and image processing, computer vision, radar and remote sensing, bioinformatics deal with high dimensional data of various types. In these applications, the high dimensional data is not generally distributed over the whole signal space; rather it lives in the union of low dimensional subspaces. Hence, classical clustering techniques depending data distributions in centroids are not successful, and techniques that facilities the low dimensional subspace structure of big data are required. Sparse subspace clustering (SSC) technique that relies on the self-expressiveness of the data is shown to provably handle the data under noiseless case for independent and disjoint subspaces. Self-expressiveness means that each data point in a union of subspaces can be efficiently represented as a linear or affine combination of data points in the set. SSC implementation involves solving an L1 minimization problem for each data point in the space and applying spectral clustering to the affinity matrix constructed by the obtained coefficients. Despite good properties, SSC suffers from high computational complexity increasing with data point numbers. In addition, for noisy data self-expressiveness does not apply anymore. This paper proposes to use perturbed orthogonal matching pursuit (POMP) within SSC framework for robust and computationally efficient estimation of the number of subspaces, their dimensions, and the segmentation of the data into each subspace. POMP was shown to be successful in recovering sparse signals under random basis perturbations, which is actually the case in corrupted data clustering. Our initial results for simulated clustering datasets show that the proposed POMP- SSC technique provides both computational efficiency and high clustering performance compared to classical SSC implementation.}, booktitle={Proceedings of SPIE - The International Society for Optical Engineering}, author={Gurbuz, A.C.}, year={2018} }
@inproceedings{gurbuz_yapici_guvenc_2018, title={Sparse channel estimation in millimeter-wave communications via parameter perturbed OMP}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85050315215&partnerID=MN8TOARS}, DOI={10.1109/ICCW.2018.8403589}, abstractNote={The millimeter-wave (mmWave) communications is a promising technology for next-generation wireless networks with its available broad spectrum. Along with massive number of antennas employed at both end of the transceiver, the number of unknown channel coefficients become extremely large. Thanks to sparse nature of mmWave links, this paper proposes a parameter perturbation based sparse recovery technique for mmWave channel estimation. Recently, classical compressive sensing (CS) based sparse recovery techniques have been applied in this area. However, CS based reconstructions are highly effected by basis mismatch problems such as off-the-grid targets, or, equivalently, scattering points. The proposed iterative algorithm called parameter perturbed orthogonal matching pursuit (PPOMP) jointly solves for both the sparse signal, which is the unknown mmWave channel itself, and the basis mismatch due to off-the-grid problem. We verify through extensive numerical results that the proposed PPOMP algorithm achieves significantly better channel estimation performance compared to the state of the art sparse reconstruction techniques.}, booktitle={2018 IEEE International Conference on Communications Workshops, ICC Workshops 2018 - Proceedings}, author={Gurbuz, A.C. and Yapici, Y. and Guvenc, I.}, year={2018}, pages={1–6} }
@article{camlica_gurbuz_arikan_2017, title={Autofocused Spotlight SAR Image Reconstruction of Off-Grid Sparse Scenes}, volume={53}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85029582577&partnerID=MN8TOARS}, DOI={10.1109/TAES.2017.2675138}, abstractNote={Synthetic aperture radar (SAR) has significant role in remote sensing. Phase errors due to uncompensated platform motion, measurement model mismatch, and measurement noise can cause degradations in SAR image reconstruction. For efficient processing of the measurements, image plane is discretized and autofocusing algorithms on this discrete grid are employed. However, in addition to the platform motion errors, the reflectors, which are not exactly on the reconstruction grid, also degrade the image quality. This is called the off-grid target problem. In this paper, a sparsity-based technique is developed for autofocused spotlight SAR image reconstruction that can correct phase errors due to uncompensated platform motion and provide robust images in the presence of off-grid targets. The proposed orthogonal matching pursuit-based reconstruction technique uses gradient descent parameter updates with built in autofocus. The technique can reconstruct high-quality images by using sub Nyquist rate of sampling on the reflected signals at the receiver. The results obtained using both simulated and real SAR system data show that the proposed technique provides higher quality reconstructions over alternative techniques in terms of commonly used performance metrics.}, number={4}, journal={IEEE Transactions on Aerospace and Electronic Systems}, author={Camlica, S. and Gurbuz, A.C. and Arikan, O.}, year={2017}, pages={1880–1892} }
@article{ilhan_gurbuz_arikan_2017, title={Compressive sensing-based robust off-the-grid stretch processing}, volume={11}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85033401459&partnerID=MN8TOARS}, DOI={10.1049/iet-rsn.2017.0133}, abstractNote={Classical stretch processing (SP) obtains high range resolution by compressing large bandwidth signals with narrowband receivers using lower rate analogue-to-digital converters. SP achieves the resolution of the large bandwidth signal by focusing into a limited range window, and by deramping in the analogue domain. SP offers moderate data rate for signal processing for high bandwidth waveforms. Furthermore, if the scene in the examined window is sparse, compressive sensing (CS)-based techniques have the potential to further decrease the required number of measurements. However, CS-based reconstructions are highly affected by model mismatches such as targets that are off-the-grid. This study proposes a sparsity-based iterative parameter perturbation technique for SP that is robust to targets off-the-grid in range or Doppler. The error between reconstructed and actual scenes is measured using Earth mover's distance metric. Performance analyses of the proposed technique are compared with classical CS and SP techniques in terms of data rate, resolution and signal-to-noise ratio. It is shown through simulations that the proposed technique offers robust and high-resolution reconstructions for the same data rate compared with both classical SP- and CS-based techniques.}, number={11}, journal={IET Radar, Sonar and Navigation}, author={Ilhan, I. and Gurbuz, A.C. and Arikan, O.}, year={2017}, pages={1730–1735} }
@article{orlando_hao_aubry_cui_gurbuz_gazor_2017, title={Special issue: advanced techniques for radar signal processing}, volume={2017}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85021656273&partnerID=MN8TOARS}, DOI={10.1186/s13634-017-0481-0}, abstractNote={Recent advances in technology have led to the development of low-cost sensing devices capable of providing high performances in terms of both computational resources and measurement precision.More important, the reduced size of such devices have allowed to exploit them in diverse application fields (as, for instance, medical, military, manufacturing, transportation, and safety systems).However, in the context of military applications (e.g., radar and communication systems), the downside of this technology is that it is available for terrorist attacks aimed at denying targeting information and using radarguided missiles or small drones carrying dangerous (e.g., explosive or chemical) substances.Thus, it stems the need for innovative signal processing solutions to counteract these threats.Such techniques are applicable in ship and aircraft monitoring (for defense purposes), coastal surveillance, and, generally speaking, homeland security.This special issue focuses on radar signal processing techniques (target detection and tracking, interference estimation and suppression, adaptive beamforming, electronic warfare) that benefit from the mentioned advances to face the new challenging operating scenarios that naturally arise from nowadays technology advantages and disadvantages.More specifically, the emphasis is on (possibly distributed) radar systems equipped with arrays of sensors, which enable to capitalize the spatial diversity and power integration enabling significant improvements in performance.In general, radar systems perform three general functions, which are search, track, and imaging.The most important operation of a search radar is target detection.As a matter of fact, once the system declares the presence of a target, its resources are scheduled to estimate target parameters [1, 2] or for target tracking [3] which consists in the fine estimation of parameters as range, azimuth angle, elevation angle, and Doppler frequency offset.}, number={1}, journal={Eurasip Journal on Advances in Signal Processing}, author={Orlando, D. and Hao, C. and Aubry, A. and Cui, G. and Gurbuz, A.C. and Gazor, S.}, year={2017} }
@inproceedings{alp_korucu_karabacak_gürbüz_arikan_2016, title={Online calibration of Modulated Wideband Converter,Kiplemeli Geniş Bant Çeviricinin Sahada Kalibrasyonu}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84982835410&partnerID=MN8TOARS}, DOI={10.1109/SIU.2016.7495889}, abstractNote={In this work, we propose a new method for online calibration of recently proposed Modulated Wideband Converter (MWC), which digitizes wideband sparse signals below the Nyquist limit without loss of information by using compressive sensing techniques. Our method requires a single frequency synthesizer card, which can generate clean tones along the operation band of the system, rather than much expensive measurement instruments such as network analyser or vector spectrum analyser, which are not appropriate for online calibration. Moreover, low computational complexity of the proposed method enables its implementation on FPGA so that it can be embedded into the system. Hence, on each power on, the system can utilize self calibration without requiring any additional measurement instruments.}, booktitle={2016 24th Signal Processing and Communication Application Conference, SIU 2016 - Proceedings}, author={Alp, Y.K. and Korucu, A.B. and Karabacak, A.T. and Gürbüz, A.C. and Arikan, O.}, year={2016}, pages={913–916} }
@inproceedings{ispir_orduyilmaz_serin_yildirim_gurbuz_2016, title={Real-Time multiple velocity false target generation in digital radio frequency memory}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84978249301&partnerID=MN8TOARS}, DOI={10.1109/RADAR.2016.7485310}, abstractNote={Real-time generation of multiple velocity false targets, which is an effective electronic counter measure (ECM) technique against pulse-doppler radar, is studied. Classical ECM techniques employ generation of velocity false targets by switching single signal source in time. This paper proposes a time and frequency domain generation approach, which is more flexible and efficient in terms of resource allocation for increased number of false targets. ECM technique is implemented in Xilinx Virtex-6 SX475t FPGA platform. Velocity false targets with different parameter sets are tested both in FPGA platform and computer environment. It is shown that the proposed technique is effective in false target generation and suitable for real time implementation requiring lower resource allocations compared to time and frequency domain generation.}, booktitle={2016 IEEE Radar Conference, RadarConf 2016}, author={Ispir, M. and Orduyilmaz, A. and Serin, M. and Yildirim, A. and Gurbuz, A.C.}, year={2016} }
@inproceedings{ortatatli_orduyilmaz_serin_özdil_yildirim_gürbüz_2016, title={Real-time frequency parameter extraction for electronic support systems,ELEKTRONIK DESTEK SISTEMLERI IÇIN GERÇEK ZAMANLI FREKANS PARAMETRESI ÇIKARIMI}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84982851440&partnerID=MN8TOARS}, DOI={10.1109/SIU.2016.7495687}, abstractNote={In this research, real time automatic frequency parameter extraction methods for electronic support systems are implemented on FPGA. The frequency parameter is extracted by using digital instantaneous frequency measurement (DIFM) and fast fourier transform (FFT) methods. Estimation performances of these two different methods for different type of radars at different signal to noise ratios (SNR) are analyzed. This design is implemented on Xilinx Virtex-6 based processing board that samples in 2.5 GHz. Experimental results of the design for FPGA implementation are presented.}, booktitle={2016 24th Signal Processing and Communication Application Conference, SIU 2016 - Proceedings}, author={Ortatatli, I.E. and Orduyilmaz, A. and Serin, M. and Özdil, O. and Yildirim, A. and Gürbüz, A.C.}, year={2016}, pages={105–108} }
@article{duman_gürbüz_2015, title={3D imaging for ground-penetrating radars via dictionary dimension reduction}, volume={23}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84938574920&partnerID=MN8TOARS}, DOI={10.3906/elk-1212-157}, abstractNote={Ground-penetrating radar (GPR) has been widely used in detecting or imaging subsurface targets. In many applications such as archaeology, utility imaging, or landmine detection, three-dimensional (3D) images of the subsurface region is required for better understanding of the sensed medium. However, a high-resolution 3D image requires wideband data collection both in spatial and time/frequency domains. Match filtering is the main tool for generating subsurface images. Applying match filtering with the data acquisition impulse response for each possible voxel in the 3D region with the collected data requires both a tremendous amount of computer memory and computational complexity. Hence, it is very costly to obtain 3D GPR images in most of the applications although 3D images are very demanded results. In this paper, a new 3D imaging technique is proposed that will first decrease the memory requirements for 3D imaging with possible implications for less computational complexity. The proposed method uses the shifted impulse response of the targets that are on the same depth as a function of scanning position. This similarity of target responses for data dictionaries for only 2D target slices is constructed with twice the length in scanning directions and this 2D dictionary is mainly used for generating 3D images. The proposed method directly saves memory due to dimension reduction in dictionary generation and also decreases computational load. Simulation results show generated 3D images with the proposed technique. Comparisons in both memory and computational load with the standard backprojection show that the proposed technique offers advantages in both areas.}, number={5}, journal={Turkish Journal of Electrical Engineering and Computer Sciences}, author={Duman, M. and Gürbüz, A.C.}, year={2015}, pages={1242–1256} }
@inproceedings{camlica_guven_gurbuz_arikan_2015, title={Analysis of sparsity based joint SAR image reconstruction and autofocus techniques}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84962833068&partnerID=MN8TOARS}, DOI={10.1109/CoSeRa.2015.7330272}, abstractNote={Synthetic Aperture Radar (SAR) has significance in many remote sensing applications. One of the main problems with SAR is the platform motion that causes defocusing in the reconstructed SAR image. To mitigate this problem, for particularly on imaging of fields that admit a sparse representation, various sparsity based techniques that either apply optimization procedures or greedy iterative solutions have been proposed in the literature. Although these techniques have been mainly compared with classical phase gradient autofocus (PGA) algorithm, they have not been analyzed and compared with each other. In this paper several of the recent sparsity based SAR phase correction techniques are compared using metrics such as mean square error (MSE), entropy, target to background ratio (TBR) in terms of undersampling ratio, signal to noise ratio (SNR). In addition to comparisons, a cross validation based stopping criterion is introduced with an OMP procedure to free the algorithm from user defined parameters. The techniques are tested on simulated data for detailed comparisons. Real data results of tested techniques are also provided. Our initial results show that all compared sparsity based techniques provide better performance compared to PGA with varying relative performances.}, booktitle={2015 3rd International Workshop on Compressed Sensing Theory and its Applications to Radar, Sonar, and Remote Sensing, CoSeRa 2015}, author={Camlica, S. and Guven, H.E. and Gurbuz, A.C. and Arikan, O.}, year={2015}, pages={99–103} }
@inproceedings{ilhan_gürbuz_2015, title={Finding sparse parametric shapes from low number of imase measurements,Seyrek Parametrik Şekillerin Görüntülerden Az Öblçüm Altlnda Tespiti}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84939168980&partnerID=MN8TOARS}, DOI={10.1109/SIU.2015.7130341}, abstractNote={Detection of parametric shapes i.e. line, circle, ellipse etc. in images is one of the most significant topics in diverse areas such as image and signal processing, pattern recognition and remote sensing. Compressive Sensing(CS) theory details how the signal is sparsely reconstructed in a known basis from low number of linear measurement. Sparsity of parametric shapes in parameter space offers to detect parametric shapes from low number of linear measurements under frameworks proposed by CS methods. Joint detedon performance of different parametric shapes in image is studied under different small number of measurements and noise level. Because of being both discrete image space and discretized parameter space, effect of offgrid, one of the most important problem in CS,is analysed in terms of shape detection. Results show that parametric shapes can robustly be found with a few measurements and effects of offgrid are seen as distribution of target energy in parameter space.}, booktitle={2015 23rd Signal Processing and Communications Applications Conference, SIU 2015 - Proceedings}, author={Ilhan, I. and Gürbuz, A.C.}, year={2015}, pages={2314–2317} }
@inproceedings{orduyilmaz_serin_yildirim_ceyhan_gürbüz_2015, title={Hybrid phase amplitude direction finding method,Hibrit Faz Genlik Yön Bulma Yöntemi}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84939157102&partnerID=MN8TOARS}, DOI={10.1109/SIU.2015.7130142}, abstractNote={In this paper, hybrid monopulse phase-amplitude comparison method is presented. Design has been realized for broadband radar signals in real time in the digital environment. Multi-antenna array is used with a wide antenna pattern disposed on a plane in direction finding (DF) techniques with a traditional phase comparison method. With recommended method in the article, integrated amplitude and phase comparison method with fractured antennas' placements is used for solution of ambiguity. It is too hard to determine phase differences by using amplitude differences coming from fractured antennas's placements. This challenge is resolved by using FFT. For a particular incidence angle sector, analysis are performed with different antennas' placement scenarios and different signal to noise ratio (SNR). Successful DF estimations are accomplished with recommended fractured, two antenna structure and hybrid phase-amplitude technique.}, booktitle={2015 23rd Signal Processing and Communications Applications Conference, SIU 2015 - Proceedings}, author={Orduyilmaz, A. and Serin, M. and Yildirim, A. and Ceyhan, K. and Gürbüz, A.C.}, year={2015}, pages={109–112} }
@inproceedings{gunyel_cinbis_ture_gurbuz_2015, title={Hyperspectral target detection - An experimental study,Hiperspektral Hedef Tespiti - Deneysel Bir Çalişma}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84939177307&partnerID=MN8TOARS}, DOI={10.1109/SIU.2015.7130427}, abstractNote={In hyperspectral imaging, the measured spectra are affected by the materials and objects that reside within or in close vicinity of the pixel which is being imaged. The detection of a material or object of interest in an imaged region is a common problem in various application areas. In this work, an experimental study is performed for target detection in hyperspectral images, supported by a performance comparison.}, booktitle={2015 23rd Signal Processing and Communications Applications Conference, SIU 2015 - Proceedings}, author={Gunyel, B. and Cinbis, R.G. and Ture, S. and Gurbuz, A.C.}, year={2015}, pages={2627–2630} }
@article{karabacak_gurbuz_gurbuz_guldogan_hendeby_gustafsson_2015, title={Knowledge Exploitation for Human Micro-Doppler Classification}, volume={12}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85027953351&partnerID=MN8TOARS}, DOI={10.1109/LGRS.2015.2452311}, abstractNote={Micro-Doppler radar signatures have great potential for classifying pedestrians and animals, as well as their motion pattern, in a variety of surveillance applications. Due to the many degrees of freedom involved, real data need to be complemented with accurate simulated radar data to be able to successfully design and test radar signal processing algorithms. In many cases, the ability to collect real data is limited by monetary and practical considerations, whereas in a simulated environment, any desired scenario may be generated. Motion capture (MOCAP) has been used in several works to simulate the human micro-Doppler signature measured by radar; however, validation of the approach has only been done based on visual comparisons of micro-Doppler signatures. This work validates and, more importantly, extends the exploitation of MOCAP data not just to simulate micro-Doppler signatures but also to use the simulated signatures as a source of a priori knowledge to improve the classification performance of real radar data, particularly in the case when the total amount of data is small.}, number={10}, journal={IEEE Geoscience and Remote Sensing Letters}, author={Karabacak, C. and Gurbuz, S.Z. and Gurbuz, A.C. and Guldogan, M.B. and Hendeby, G. and Gustafsson, F.}, year={2015}, pages={2125–2129} }
@inproceedings{aydemir_gurbuz_bahceci_2015, title={Non-Linear junction detectors: Experimental performance analysis,Dogrusal Olmayan Kavşak Detektörleri: Deneysel Performans Analizi}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84939138727&partnerID=MN8TOARS}, DOI={10.1109/SIU.2015.7130332}, abstractNote={Non-linear junction detectors (NLJD) are used for detecting objects which contain p-n junction or corrosive materials (m-m junction) in a target area. p-njunctions are mostly included in electronic circuits which may contain diodes, transistors, amplifiers etc. Typical NLJDs employ harmonic radar principels where the 2 nd and 3 rd harmonics created by the target area after illimunation by a sinusoidal signal are compared to distinguish whether a p-n or m-m junction or nothing at all exists in the area. In general, the p-n junctions create stronger 2 nd harmonics than the m-m junctions do. The performance of NLJD is measured over how well it can differentiate p-n and m-m junctions. In this paper, we investiagted the factors that effect the NLJD performance. It is observed that the single tone illuminater signal has inferior performance. Therefore, we investigated a 2-tone illuminator signal and analyzed the impact of additoal harmoncis created by the combination of sums and differences of the two illuminating tones.}, booktitle={2015 23rd Signal Processing and Communications Applications Conference, SIU 2015 - Proceedings}, author={Aydemir, C. and Gurbuz, A.C. and Bahceci, I.}, year={2015}, pages={2278–2281} }
@inproceedings{orduyilmaz_kara_serin_yildirim_gürbüz_efe_2015, title={Real-time pulse compression radar waveform generation and digital matched filtering}, volume={2015-June}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84937903849&partnerID=MN8TOARS}, DOI={10.1109/RADAR.2015.7131037}, abstractNote={Real-time digital implementation of radar waveform pulse compression and match filtering on FPGA platform is studied. In this work, different types of radar waveforms including phase coded, linear frequency (LFM) and non-linear frequency modulated (NLFM) signals are generated digitally in Xilinx Virtex-5 FPGA platform. Waveforms with different time bandwidth products are tested both in FPGA platform and computer. Digital matched filtering implementation procedure used in FPGA is presented and comparison of theoretical calculations and FPGA implementation results along with implementation resource utilization are presented. Results indicate that precise generation of real-time waveform matched filtering implementations deviate at most 1 dB on range sidelobe levels from theoretical results. Moreover adopted segmentation and parallel implementation of the received pulse both allows processing of divided pulses without SNR degradation and uses less FPGA resources in general compared to processing full PRI at once.}, number={June}, booktitle={IEEE National Radar Conference - Proceedings}, author={Orduyilmaz, A. and Kara, G. and Serin, M. and Yildirim, A. and Gürbüz, A.C. and Efe, M.}, year={2015}, pages={426–431} }
@inproceedings{camlica_gurbuz_arikan_2015, title={SAR image reconstruction with joint off-grid target and phase error corrections}, volume={2015-November}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84962592160&partnerID=MN8TOARS}, DOI={10.1109/IGARSS.2015.7326828}, abstractNote={Synthetic Aperture Radar (SAR) has significant importance in many remote sensing applications. Errors due to the platform motion or measurement model uncertainties can cause degradations in the constructed SAR images. Most of the methods deal with the phase errors which cause defocusing on the image. For efficient processing of the measurements, they discretize the fast time-slow time plane and then employ autofocus algorithms on this discrete grid. However, the reflectors which are not placed exactly on the grid degrade the image quality considerably. This is the most probable case in the practical SAR operation and it causes blur or spark like affects on the image. This is called the off-grid target problem. In this work, a Compressed Sensing based technique is developed which constructs spotlight mode SAR image, handles the off-grid target problem and makes autofocus simultaneously. A gradient descent type iterative solution is used.}, booktitle={International Geoscience and Remote Sensing Symposium (IGARSS)}, author={Camlica, S. and Gurbuz, A.C. and Arikan, O.}, year={2015}, pages={4502–4505} }
@inproceedings{ilhan_gurbuz_arikan_2015, title={Sparsity based robust Stretch Processing}, volume={2015-September}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84961386544&partnerID=MN8TOARS}, DOI={10.1109/ICDSP.2015.7251837}, abstractNote={Strecth Processing (SP) is a radar signal processing technique that provides high-range resolution with processing large bandwidth signals with lower rate Analog to Digital Converter(ADC)s. The range resolution of the large bandwidth signal is obtained through looking into a limited range window and low rate ADC samples. The target space in the observed range window is sparse and Compressive sensing(CS) is an important tool to further decrease the number of measurements and sparsely reconstruct the target space for sparse scenes with a known basis which is the Fourier basis in the general application of SP. Although classical CS techniques might be directly applied to SP, due to off-grid targets reconstruction performance degrades. In this paper, applicability of compressive sensing framework and its sparse signal recovery techniques to stretch processing is studied considering off-grid cases. For sparsity based robust SP, Perturbed Parameter Orthogonal Matching Pursuit(PPOMP) algorithm is proposed. PPOMP is an iterative technique that estimates off-grid target parameters through a gradient descent. To compute the error between actual and reconstructed parameters, Earth Movers Distance(EMD) is used. Performance of proposed algorithm are compared with classical CS and SP techniques.}, booktitle={International Conference on Digital Signal Processing, DSP}, author={Ilhan, I. and Gurbuz, A.C. and Arikan, O.}, year={2015}, pages={95–99} }
@inproceedings{teke_arikan_gurbuz_2014, title={A recursive approach to reconstruction of sparse signals,Seyrek sinyallerin geri çatimina özyineli bir yaklaşim}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84903765587&partnerID=MN8TOARS}, DOI={10.1109/SIU.2014.6830436}, abstractNote={Compressive Sensing (CS) theory details how a sparsely represented signal in a known basis can be reconstructed using less number of measurements. In many practical systems, the observation signal has a sparse representation in a continuous parameter space. This situation rises the possibility of use of the CS reconstruction techniques in the practical problems. In order to utilize CS techniques, the continuous parameter space have to be discretized. This discritization brings the well-known off-grid problem. To prevent the off-grid problem, this study offers a recursive approach which discritizes the parameter space in an adaptive manner. The simulations show that the proposed approach can estimate the parameters with a high accuracy even if targets are closely spaced.}, booktitle={2014 22nd Signal Processing and Communications Applications Conference, SIU 2014 - Proceedings}, author={Teke, O. and Arikan, O. and Gurbuz, A.C.}, year={2014}, pages={1142–1145} }
@inproceedings{teke_gurbuz_arikan_2015, title={A recursive way for sparse reconstruction of parametric spaces}, volume={2015-April}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84940557695&partnerID=MN8TOARS}, DOI={10.1109/ACSSC.2014.7094524}, abstractNote={A novel recursive framework for sparse reconstruction of continuous parameter spaces is proposed by adaptive partitioning and discretization of the parameter space together with expectation maximization type iterations. Any sparse solver or reconstruction technique can be used within the proposed recursive framework. Experimental results show that proposed technique improves the parameter estimation performance of classical sparse solvers while achieving Cramér-Rao lower bound on the tested frequency estimation problem.}, booktitle={Conference Record - Asilomar Conference on Signals, Systems and Computers}, author={Teke, O. and Gurbuz, A.C. and Arikan, O.}, year={2015}, pages={637–641} }
@article{teke_gurbuz_arikan_2014, title={A robust compressive sensing based technique for reconstruction of sparse radar scenes}, volume={27}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84897632309&partnerID=MN8TOARS}, DOI={10.1016/j.dsp.2013.12.008}, number={1}, journal={Digital Signal Processing: A Review Journal}, author={Teke, O. and Gurbuz, A.C. and Arikan, O.}, year={2014}, pages={23–32} }
@inproceedings{analysis of frequency modulated continuous wave signals using time-frequency domain shape features,zaman-frekans düzlemi şekil özellikleri kullanilarak sürekli dalga frekans modülasyonlu işaret analizi_2014, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84903763793&partnerID=MN8TOARS}, DOI={10.1109/SIU.2014.6830595}, abstractNote={In this paper, modulation parameters as modulation bandwidth and type are extracted by using the shape features of the time-frequency domain of the frequency modulated continuous wave radar (FMCW) signals. Time-frequency image of the received signal is obtained by taking short time fourier transform (STFT). Then Hough transform (HT) is applied to the binary image which is obtained after thresholding. Modulation parameters are determined after the evaluation of the one dimensional modulation vector. The result of the parameter extraction at different SNR values and modulation types is presented.}, booktitle={2014 22nd Signal Processing and Communications Applications Conference, SIU 2014 - Proceedings}, year={2014}, pages={1778–1781} }
@inproceedings{karabacak_gurbuz_gurbuz_2014, title={Automatic human activity classification using radar,Radar kullanarak farkli insan hareketlerinin otomatik siniflandirilmasi}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84903766310&partnerID=MN8TOARS}, DOI={10.1109/SIU.2014.6830413}, abstractNote={Developing automatic target classification algorithms using radar is a widely researched topic in recent years. Among the targets classified in these algorithms, the most studied is humans. Classification of humans with high accuracy is very significant for many military and civil applications. Moreover, in these studies, besides the classification of human target, activities are also analyzed. Knowledge of the activity a person is engaged in can substantially change the alarm level in some applications. In this paper, an algorithm that automatically classifies walking, running, crawling, and creeping using radar is presented.}, booktitle={2014 22nd Signal Processing and Communications Applications Conference, SIU 2014 - Proceedings}, author={Karabacak, C. and Gurbuz, S.Z. and Gurbuz, A.C.}, year={2014}, pages={1051–1054} }
@inproceedings{albayrak_gurbuz_gunyel_2014, title={Compressed sensing based hyperspectral unmixing,Sikiştirilmiş algilama ile hiperspektral ayriştirma}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84903761087&partnerID=MN8TOARS}, DOI={10.1109/SIU.2014.6830510}, abstractNote={In hyperspectral images the measured spectra for each pixel can be modeled as convex combination of small number of endmember spectra. Since the measured structure contains only a few of possible responses out of possibly many materials sparsity based convex optimization techniques or compressive sensing can be used for hyperspectral unmixing. In this work varying sparsity based techniques are tested for hyperspectral unmixing problem. Performance analysis of these techniques on sparsity level and measurement number are performed. Effect of high coherence of hyperspectral dictionaries is disccussed and effect of signal to noise ratio is analyzed.}, booktitle={2014 22nd Signal Processing and Communications Applications Conference, SIU 2014 - Proceedings}, author={Albayrak, R.T. and Gurbuz, A.C. and Gunyel, B.}, year={2014}, pages={1438–1441} }
@inproceedings{orduyilmaz_serin_gurbuz_yildirim_2014, title={Passive direction finding using amplitude and phase comparison techniques,Genlik ve faz karşilaştirma yöntemi ile pasif yön bulma}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84903793857&partnerID=MN8TOARS}, DOI={10.1109/SIU.2014.6830384}, abstractNote={Amplitude and phase comparison methods are practical direction finding (DF) techniques for electronic support and electronic intelligence systems. The combination of amplitude and phase comparison methods improves the direction finding accuracy and at the same time the ambiguities of phase comparison method are resolved by amplitude comparison method. In this research, the important factors that affect the accuracy of direction finding algorithm, such as the number of antennas, antennas' placements and distances between antennas are analyzed with different signal to noise ratio (SNR) values. The root mean square (RMS) error values of direction of arrival estimates are presented and the antenna placement scenarios for high DF accuracy are acquired.}, booktitle={2014 22nd Signal Processing and Communications Applications Conference, SIU 2014 - Proceedings}, author={Orduyilmaz, A. and Serin, M. and Gurbuz, A.C. and Yildirim, A.}, year={2014}, pages={935–938} }
@inproceedings{eroi_karabacak_gurbuz_gurbuz_2014, title={Radar simulation of different human activities via Kinect,Kinect ile farkli insan hareketlerinin radar benzetimi}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84903766486&partnerID=MN8TOARS}, DOI={10.1109/SIU.2014.6830404}, abstractNote={The classification of different human activities with radar has been a widely researched topic in recent years. Oftentimes, when no experimental data is available, simulated data can be exploited to test classification algorithms. Kinematic models such as the Thalmann Model and motion capture (MOCAP) data are frequently used to simulate radar signatures of human movements. While the Thalmann Model provides a model only for human walking, MOCAP data has the capability to supply data for almost any type of human activity. However, most commercial MOCAP data acquisition systems are quite expensive, making it difficult to obtain MOCAP data. In this paper, economical, easily obtainable and practical Kinect sensor is used to develop a skeleton tracking algorithm. In this way, simulated radar micro-Doppler signatures for different people and activities are computed.}, booktitle={2014 22nd Signal Processing and Communications Applications Conference, SIU 2014 - Proceedings}, author={Eroi, B. and Karabacak, C. and Gurbuz, S.Z. and Gurbuz, A.C.}, year={2014}, pages={1015–1018} }
@article{ugur_arikan_gürbüz_2015, title={SAR image reconstruction by expectation maximization based matching pursuit}, volume={37}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84922537289&partnerID=MN8TOARS}, DOI={10.1016/j.dsp.2014.11.001}, number={1}, journal={Digital Signal Processing: A Review Journal}, author={Ugur, S. and Arikan, O. and Gürbüz, A.C.}, year={2015}, pages={75–84} }
@inproceedings{erol_karabacak_gurbuz_gurbuz_2014, title={Simulation of human micro-Doppler signatures with Kinect sensor}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84906751144&partnerID=MN8TOARS}, DOI={10.1109/RADAR.2014.6875712}, abstractNote={The availability and access to real radar data collected for targets with a desired characteristic is often limited by monetary and practical resources, especially in the case of airborne radar. In such cases, the generation of accurate simulated radar data is critical to the successful design and testing of radar signal processing algorithms. In the case of human micro-Doppler research, simulations of the expected target signature are required for a wide parameter space, including height, weight, gender, range, angle and waveform. The applicability of kinematic models is limited to just walking, while the use of motion capture databases is restricted to the test subjects and scenarios recorded by a third-party. To enable the simulation of human micro-Doppler signatures at will, this work exploits the inexpensive Kinect sensor to generate human spectrograms of any motion and for any subject from skeleton tracking data. The simulated spectrograms generated are statistically compared with those generated from high quality motion capture data. It is shown that the Kinect spectrograms are of sufficient quality to be used in simulation and classification of human micro-Doppler.}, booktitle={IEEE National Radar Conference - Proceedings}, author={Erol, B. and Karabacak, C. and Gurbuz, S.Z. and Gurbuz, A.C.}, year={2014}, pages={863–868} }
@inproceedings{teke_gurbuz_arikan_2014, title={Sparse delay-Doppler image reconstruction under off-grid problem}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84907411964&partnerID=MN8TOARS}, DOI={10.1109/SAM.2014.6882429}, abstractNote={Pulse-Doppler radar has been successfully applied to surveillance and tracking of both moving and stationary targets. For efficient processing of radar returns, delay-Doppler plane is discretized and FFT techniques are employed to compute matched filter output on this discrete grid. However, for targets whose delay-Doppler values do not coincide with the computation grid, the detection performance degrades considerably. Especially for detecting strong and closely spaced targets this causes miss detections and false alarms. Although compressive sensing based techniques provide sparse and high resolution results at sub-Nyquist sampling rates, straightforward application of these techniques is significantly more sensitive to the off-grid problem. Here a novel and OMP based sparse reconstruction technique with parameter perturbation, named as PPOMP, is proposed for robust delay-Doppler radar processing even under the off-grid case. In the proposed technique, the selected dictionary parameters are perturbed towards directions to decrease the orthogonal residual norm. A new performance metric based on Kull-back-Leibler Divergence (KLD) is proposed to better characterize the error between actual and reconstructed parameter spaces.}, booktitle={Proceedings of the IEEE Sensor Array and Multichannel Signal Processing Workshop}, author={Teke, O. and Gurbuz, A.C. and Arikan, O.}, year={2014}, pages={409–412} }
@article{karakus_gurbuz_tavli_2013, title={Analysis of energy efficiency of compressive sensing in wireless sensor networks}, volume={13}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84876221016&partnerID=MN8TOARS}, DOI={10.1109/JSEN.2013.2244036}, abstractNote={Improving the lifetime of wireless sensor networks (WSNs) is directly related to the energy efficiency of computation and communication operations in the sensor nodes. Compressive sensing (CS) theory suggests a new way of sensing the signal with a much lower number of linear measurements as compared to the conventional case provided that the underlying signal is sparse. This result has implications on WSN energy efficiency and prolonging network lifetime. In this paper, the effects of acquiring, processing, and communicating CS-based measurements on WSN lifetime are analyzed in comparison to conventional approaches. Energy dissipation models for both CS and conventional approaches are built and used to construct a mixed integer programming framework that jointly captures the energy costs for computation and communication for both CS and conventional approaches. Numerical analysis is performed by systematically sampling the parameter space (i.e., sparsity levels, network radius, and number of nodes). Our results show that CS prolongs network lifetime for sparse signals and is more advantageous for WSNs with a smaller coverage area.}, number={5}, journal={IEEE Sensors Journal}, author={Karakus, C. and Gurbuz, A.C. and Tavli, B.}, year={2013}, pages={1999–2008} }
@inproceedings{tekeli_gurbuz_yuksel_gurbuz_guldogan_2013, title={Classification of human micro-Doppler in a radar network}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84884823239&partnerID=MN8TOARS}, DOI={10.1109/RADAR.2013.6586080}, abstractNote={The unique, bi-pedal motion of humans has been shown to generate a characteristic micro-Doppler signature in the time-frequency domain that can be used to discriminate humans from not just other targets, but also between different activities, such as walking and running. However, the classification performance increasingly drops as the aspect angle between the target and radar approaches perpendicular, and the radial velocity component seen by the radar is minimized. In this paper, exploitation of the multi-static micro-Doppler signature formed from multi-angle observations of a radar network is proposed to improve oblique-angle classification performance. The concept of mutual information is applied to find the order of importance of features for a given classification problem, thereby enabling the selection of optimal features prior to classification. Strategies for fusing multistatic data using mutual information and model-based approaches are discussed.}, booktitle={IEEE National Radar Conference - Proceedings}, author={Tekeli, B. and Gurbuz, S.Z. and Yuksel, M. and Gurbuz, A.C. and Guldogan, M.B.}, year={2013} }
@inproceedings{teke_arikan_gürbüz_2013, title={Compressive sensing based target detection in delay-doppler radars,Sikiştirilmiş algilama ile darbe-doppler radar hedef tespiti}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84880880418&partnerID=MN8TOARS}, DOI={10.1109/SIU.2013.6531496}, abstractNote={Compressive Sensing theory shows that, a sparse signal can be reconstructed from its sub-Nyquist rate random samples. With this property, CS approach has many applications. Radar systems, which deal with sparse signal due to its nature, is one of the important application of CS theory. Even if CS approach is suitable for radar systems, classical detections schemes under Neyman-Pearson formulations may result high probability of false alarm, when CS approach is used, especially if the target has off-grid parameters. In this study, a new detection scheme which enables CS techniques to be used in radar systems is investigated.}, booktitle={2013 21st Signal Processing and Communications Applications Conference, SIU 2013}, author={Teke, O. and Arikan, O. and Gürbüz, A.C.}, year={2013} }
@inproceedings{ataman_tekeli_gürbüz_2013, title={Development of a stepped frequency GPR prototype,Basamak frekans yapida GPR prototipi geliştirilmesi}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84880907110&partnerID=MN8TOARS}, DOI={10.1109/SIU.2013.6531413}, abstractNote={Ground penetrating radar systems and through the wall radars are important research areas which are frequently studied on due to their civilian and military applications. Development of penetrating radar systems using microwave frequencies is primary in that point. In this work, a stepped frequency continuous wave (SFCW) GPR prototype is developed. The system is working between 700MHz–6GHz frequency range and is controlled through user interface. The measured data and the detection results are shown through the same user interface in real time. The system can be used either for through the wall or subsurface applications. Initial experimental data results are shown.}, booktitle={2013 21st Signal Processing and Communications Applications Conference, SIU 2013}, author={Ataman, A. and Tekeli, B. and Gürbüz, A.C.}, year={2013} }
@inproceedings{gurbuz_tekeli_yuksel_karabacak_gurbuz_guldogan_2013, title={Importance ranking of features for human micro-Doppler classification with a radar network}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84890824413&partnerID=MN8TOARS}, booktitle={Proceedings of the 16th International Conference on Information Fusion, FUSION 2013}, author={Gurbuz, S.Z. and Tekeli, B. and Yuksel, M. and Karabacak, C. and Gurbuz, A.C. and Guldogan, M.B.}, year={2013}, pages={610–616} }
@inproceedings{karabacak_gürbüz_guldogan_gürbüz_2013, title={Multi-aspect angle classification of human radar signatures}, volume={8734}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84881143727&partnerID=MN8TOARS}, DOI={10.1117/12.2017709}, abstractNote={The human micro-Doppler signature is a unique signature caused by the time-varying motion of each point on the human body, which can be used to discriminate humans from other targets exhibiting micro-Doppler, such as vehicles, tanks, helicopters, and even other animals. Classification of targets based on micro-Doppler generally involves joint timefrequency analysis of the radar return coupled with extraction of features that may be used to identify the target. Although many techniques have been investigated, including artificial neural networks and support vector machines, almost all suffer a drastic drop in classification performance as the aspect angle of human motion relative to the radar increases. This paper focuses on the use of radar networks to obtain multi-aspect angle data and thereby ameliorate the dependence of classification performance on aspect angle. Knowledge of human walking kinematics is exploited to generate a fuse spectrogram that incorporates estimates of model parameters obtained from each radar in the network. It is shown that the fused spectrogram better approximates the truly underlying motion of the target observed as compared with spectrograms generated from individual nodes.}, booktitle={Proceedings of SPIE - The International Society for Optical Engineering}, author={Karabacak, C. and Gürbüz, S.Z. and Guldogan, M.B. and Gürbüz, A.C.}, year={2013} }
@inproceedings{ugur_arikan_gurbuz_2013, title={Off-grid sparse SAR image reconstruction by EMMP algorithm}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84884847477&partnerID=MN8TOARS}, DOI={10.1109/RADAR.2013.6586034}, abstractNote={A new and robust sparse SAR image reconstruction technique is proposed for off-grid targets in the CS framework. In the proposed approach, basis vectors corresponding to on-grid point reflectors are perturbed on a finer grid to find the appropriate bases for the reconstruction of off-grid targets. To provide efficiency of the reconstruction, the EMMP algorithm is applied to find reflectivity center locations. As demonstrated by simulations, the proposed approach significantly improves the performance of sparse SAR image reconstruction.}, booktitle={IEEE National Radar Conference - Proceedings}, author={Ugur, S. and Arikan, O. and Gurbuz, A.C.}, year={2013} }
@article{teke_gurbuz_arikan_2013, title={Perturbed orthogonal matching pursuit}, volume={61}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84888413209&partnerID=MN8TOARS}, DOI={10.1109/TSP.2013.2283840}, abstractNote={Compressive Sensing theory details how a sparsely represented signal in a known basis can be reconstructed with an underdetermined linear measurement model. However, in reality there is a mismatch between the assumed and the actual bases due to factors such as discretization of the parameter space defining basis components, sampling jitter in A/D conversion, and model errors. Due to this mismatch, a signal may not be sparse in the assumed basis, which causes significant performance degradation in sparse reconstruction algorithms. To eliminate the mismatch problem, this paper presents a novel perturbed orthogonal matching pursuit (POMP) algorithm that performs controlled perturbation of selected support vectors to decrease the orthogonal residual at each iteration. Based on detailed mathematical analysis, conditions for successful reconstruction are derived. Simulations show that robust results with much smaller reconstruction errors in the case of perturbed bases can be obtained as compared to standard sparse reconstruction techniques.}, number={24}, journal={IEEE Transactions on Signal Processing}, author={Teke, O. and Gurbuz, A.C. and Arikan, O.}, year={2013}, pages={6220–6231} }
@inproceedings{karabacak_gürbüz_gürbüz_2013, title={Radar simulation of human micro-doppler signature from video motion capture data,Video hareket yakalama verilerini kullanarak insan mikro-doppler imzasinin radar simülasyonu}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84880898181&partnerID=MN8TOARS}, DOI={10.1109/SIU.2013.6531365}, abstractNote={The availability of data sets on which signal processing techniques may be tested is critical to the development of human detection, identification, and classification algorithms. However, in many cases real radar data of the desired characteristics may be expensive or difficult to obtain. In this case, synthetic or simulated data is desired. Much of the simulated data used in publications is derived from the Boulic kinematic model. But, the Boulic model is only valid for walking and is not applicable to compute the micro-Doppler signatures of other human motions. The Carnegie Mellon University motion capture library includes data from a wide range of human activities and provides the time-varying position of body parts. In this work, this video motion capture data is used to generate the radar micro-Doppler signature for many human activities. Observations about the micro-Doppler signatures computed are also shared.}, booktitle={2013 21st Signal Processing and Communications Applications Conference, SIU 2013}, author={Karabacak, C. and Gürbüz, S.Z. and Gürbüz, A.C.}, year={2013} }
@article{gurbuz_teke_arikan_2013, title={Sparse ground-penetrating radar imaging method for off-the-grid target problem}, volume={22}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84892710786&partnerID=MN8TOARS}, DOI={10.1117/1.JEI.22.2.021007}, abstractNote={Spatial sparsity of the target space in subsurface or through-the-wall imaging applications has been successfully used within the compressive-sensing framework to decrease the data acquisition load in practical systems, while also generating high-resolution images. The developed techniques in this area mainly discretize the continuous target space into grid points and generate a dictionary of model data that is used in image-reconstructing optimization problems. However, for targets that do not coincide with the computation grid, imaging performance degrades considerably. This phenomenon is known as the off-grid problem. This paper presents a novel sparse ground-penetrating radar imaging method that is robust for off-grid targets. The proposed technique is an iterative orthogonal matching pursuit-based method that uses gradient-based steepest ascent-type iterations to locate the off-grid target. Simulations show that robust results with much smaller reconstruction errors are obtained for multiple off-grid targets compared to standard sparse reconstruction techniques.}, number={2}, journal={Journal of Electronic Imaging}, author={Gurbuz, A.C. and Teke, O. and Arikan, O.}, year={2013} }
@inproceedings{teke_gürbüz_arikan_2012, title={A new OMP technique for sparse recovery,Seyrek geriçatma i̇çin yeni bir OMP yöntemi}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84863468198&partnerID=MN8TOARS}, DOI={10.1109/SIU.2012.6204606}, abstractNote={Compressive Sensing (CS) theory details how a sparsely represented signal in a known basis can be reconstructed using less number of measurements. However in reality there is a mismatch between the assumed and the actual bases due to several reasons like discritization of the parameter space or model errors. Due to this mismatch, a sparse signal in the actual basis is definitely not sparse in the assumed basis and current sparse reconstruction algorithms suffer performance degradation. This paper presents a novel orthogonal matching pursuit algorithm that has a controlled perturbation mechanism on the basis vectors, decreasing the residual norm at each iteration. Superior performance of the proposed technique is shown in detailed simulations.}, booktitle={2012 20th Signal Processing and Communications Applications Conference, SIU 2012, Proceedings}, author={Teke, O. and Gürbüz, A.C. and Arikan, O.}, year={2012} }
@inproceedings{duman_gurbuz_2012, title={Analysis of compressive sensing based through the wall imaging}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84864228336&partnerID=MN8TOARS}, DOI={10.1109/RADAR.2012.6212218}, abstractNote={Compressive sensing (CS) has been shown to be a useful tool for subsurface or through the wall imaging (TWI) using ground penetrating radar (GPR). It has been used to decrease both time/frequency or spatial measurements or to generate high resolution images. Although current works apply CS to TWI, they lack analysis for CS about the required number of measurements for sparsity levels, imaging performance in varying noise levels or performance of different measurement strategies. In addition proposed CS based imaging methods are based on two basic assumptions; targets are point like positioned at only discrete spatial or grid locations and wall thickness and its dielectric constant are perfectly known. However these assumptions are not usually valid in most TWI applications. This work details the theory for CS based TWI, analyzes the performance of the proposed imaging for the above mentioned cases. The effect of errors in unknown parameters on the imaging performance is analyzed and possible solutions are discussed.}, booktitle={IEEE National Radar Conference - Proceedings}, author={Duman, M. and Gurbuz, A.C.}, year={2012}, pages={0641–0646} }
@article{tuncer_gürbüz_2012, title={Analysis of orthogonal matching pursuit based subsurface imaging for compressive ground penetrating radars}, volume={20}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84865270621&partnerID=MN8TOARS}, DOI={10.3906/elk-1104-3}, abstractNote={It is shown that compressive sensing (CS) theory can be used for subsurface imaging in stepped frequency ground penetrating radars (GPR), resulting in robust sparse images, using fewer measurements. Although the data acquisition time is decreased by CS, the computational complexity of the minimization based imaging algorithm is too costly, which makes the algorithm useless, especially for extensive discretization or 3D imaging. In this paper, a greedy alternative, orthogonal matching pursuit (OMP) is used for imaging subsurface and its performance under various conditions is compared to CS imaging method. Results show that OMP could reconstruct sparse signals robustly as well as CS imaging. It is faster and easier to implement so it can be said that OMP is a fascinating alternative to CS imaging method for subsurface GPR imaging.}, number={6}, journal={Turkish Journal of Electrical Engineering and Computer Sciences}, author={Tuncer, M.A.C. and Gürbüz, A.C.}, year={2012}, pages={979–989} }
@article{gurbuz_cevher_mcclellan_2012, title={Bearing estimation via spatial sparsity using compressive sensing}, volume={48}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84859855594&partnerID=MN8TOARS}, DOI={10.1109/TAES.2012.6178067}, abstractNote={Bearing estimation algorithms obtain only a small number of direction of arrivals (DOAs) within the entire angle domain, when the sources are spatially sparse. Hence, we propose a method to specifically exploit this spatial sparsity property. The method uses a very small number of measurements in the form of random projections of the sensor data along with one full waveform recording at one of the sensors. A basis pursuit strategy is used to formulate the problem by representing the measurements in an overcomplete dictionary. Sparsity is enforced by $\ell_1$ -norm minimization which leads to a convex optimization problem that can be efficiently solved with a linear program. This formulation is very effective for decreasing communication loads in multi sensor systems. The algorithm provides increased bearing resolution and is applicable for both narrowband and wideband signals. Sensors positions must be known, but the array shape can be arbitrary. Simulations and field data results are provided to demonstrate the performance and advantages of the proposed method.}, number={2}, journal={IEEE Transactions on Aerospace and Electronic Systems}, author={Gurbuz, A.C. and Cevher, V. and McClellan, J.H.}, year={2012}, pages={1358–1369} }
@inproceedings{ilbe?i_gürbüz_2012, title={Demosaicking with compressive sensing,Sikiştirilmiş algilama i̇le demozai̇kleme}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84863477371&partnerID=MN8TOARS}, DOI={10.1109/SIU.2012.6204758}, abstractNote={Sparse signals can be recovered with less number of measurements compared to standard methods using Compressive Sensing (CS) theory. In digital cameras, color filter arrays (CFA) are used to sample each color band with less measurements than the normal. The color images are reconstructed using interpolation of measured pixel values. In this study, assuming images are sparse or compressible in a basis demosaicking is done with CS using the measurements from the CFA pattern. Separate, together and joint sparsity models are used for reconstructing images. Reconstructed sparsity levels for used CFA patterns are found. The images reconstructed with the proposed method are compared with the results from bilinear interpolation.}, booktitle={2012 20th Signal Processing and Communications Applications Conference, SIU 2012, Proceedings}, author={Ilbe?i, H. and Gürbüz, A.C.}, year={2012} }
@inproceedings{karaku?_gürbüz_tavli_2012, title={Efficiency of compressive sensing on the lifetime of wireless sensor networks,Sikiştirilmiş algilamanin kablosuz algilayici aǧlarin yaşam süresi̇ üzeri̇ndeki̇ etki̇nli̇ ǧi̇}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84863490220&partnerID=MN8TOARS}, DOI={10.1109/SIU.2012.6204564}, abstractNote={Improving the lifetime of Wireless Sensor Networks (WSNs) is directly related with the energy efficiency of computation and communication operations in the sensor nodes. By employing the concepts of Compressive Sensing (CS) theory it is possible to reconstruct a signal with a certain number of random linear measurements, which is much less than the number of measurements necessary in conventional signal reconstruction techniques. In this study, communication and computation energy dissipation models for CS and conventional signal processing techniques are built in order to investigate impact of CS based signal reconstruction on WSN lifetime with Linear Programming (LP) based on the data flow optimization.}, booktitle={2012 20th Signal Processing and Communications Applications Conference, SIU 2012, Proceedings}, author={Karaku?, C. and Gürbüz, A.C. and Tavli, B.}, year={2012} }
@inproceedings{gurbuz_pilanci_arikan_2012, title={Expectation maximization based matching pursuit}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84867589179&partnerID=MN8TOARS}, DOI={10.1109/ICASSP.2012.6288624}, abstractNote={A novel expectation maximization based matching pursuit (EMMP) algorithm is presented. The method uses the measurements as the incomplete data and obtain the complete data which corresponds to the sparse solution using an iterative EM based framework. In standard greedy methods such as matching pursuit or orthogonal matching pursuit a selected atom can not be changed during the course of the algorithm even if the signal doesn't have a support on that atom. The proposed EMMP algorithm is also flexible in that sense. The results show that the proposed method has lower reconstruction errors compared to other greedy algorithms using the same conditions.}, booktitle={ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings}, author={Gurbuz, A.C. and Pilanci, M. and Arikan, O.}, year={2012}, pages={3313–3316} }
@article{duman_gurbuz_2012, title={Performance analysis of compressive-sensing-based through-the-wall imaging with effect of unknown parameters}, volume={2012}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84864976354&partnerID=MN8TOARS}, DOI={10.1155/2012/405145}, abstractNote={Compressive sensing (CS) has been shown to be a useful tool for subsurface or through-the-wall imaging (TWI) using ground penetrating radar (GPR). It has been used to decrease both time/frequency or spatial measurements and generate high-resolution images. Although current works apply CS directly to TWI, questions on the required number of measurements for a sparsity level, measurement strategy to subsample in frequency and space, or imaging performance in varying noise levels and limits on CS range resolution performance still needs to be answered. In addition current CS-based imaging methods are based on two basic assumptions; targets are point like and positioned at only discrete grid locations and wall thickness and its dielectric constant are perfectly known. However, these assumptions are not usually valid in most TWI applications. This work extends the theory of CS-based radar imaging developed for subsurface imaging to TWI and outlines the performance of the proposed imaging for the above-mentioned questions using numerical simulations. The effect of unknown parameters on the imaging performance is analyzed, and it is observed that off-the-grid point targets and big modeling errors decreases the performance of CS imaging.}, journal={International Journal of Antennas and Propagation}, author={Duman, M. and Gurbuz, A.C.}, year={2012} }
@inproceedings{akta?_gürbüz_2012, title={Recognition of stagger pulse repetition interval for electronic support systems,Elektroni̇k destek si̇stemleri̇ i̇çi̇n atlamali darbe tekrar araliǧinin algilanmasi}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84863454566&partnerID=MN8TOARS}, DOI={10.1109/SIU.2012.6204510}, abstractNote={Stagger Pulse Repetition Interval (PRI) modulation has been widely used in radar systems. Stagger PRI modulation is hard to recognize because it has the same characteristics with jitter PRI modulation. Electronic support systems have to work with spurious and missing pulses due to hard environment conditions. This situation also makes it difficult to recognize the stagger PRI modulated signals. On the other hand, since response time of the system is also very important for the electronic support systems, the designed method should be simple and causal. In this paper we propose a new causal recognition method ,that does not use histogram based techniques, for stagger PRI modulation in dense environment.}, booktitle={2012 20th Signal Processing and Communications Applications Conference, SIU 2012, Proceedings}, author={Akta?, M.C. and Gürbüz, A.C.}, year={2012} }
@inproceedings{u?ur_arikan_gürbüz_2012, title={SAR image reconstruction by EMMP algorithm,EMMP algori̇tmasi kullanilarak SAR görüntü oluşturulmasi}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84863455166&partnerID=MN8TOARS}, DOI={10.1109/SIU.2012.6204604}, abstractNote={In this work, EMMP algorithm is used to solve the SAR image reconstruction problem which is modelled in the compressed sensing context. It is found that the sparsity parameter of the target region is an important parameter determining the quality of the output image. The proposed method is applied to the real SAR data and provided high quality outputs.}, booktitle={2012 20th Signal Processing and Communications Applications Conference, SIU 2012, Proceedings}, author={U?ur, S. and Arikan, O. and Gürbüz, A.C.}, year={2012} }
@inproceedings{duman_gürbüz_2012, title={Through the wall imaging with compressive sensing and effects of unknown parameters to the performance,Sikiştirilmiş algilama i̇le duvar arkasi görüntüleme ve bi̇li̇nmeyen parametreleri̇n performansa etki̇leri̇}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84863463124&partnerID=MN8TOARS}, DOI={10.1109/SIU.2012.6204473}, abstractNote={Compressive Sensing (CS) is a very useful algorithm that is used in Ground Penetrating Radar (GPR) or through the wall imaging (TWI). CS generates high resolution images by decreasing the required measurement number. Although there are several recent studies about TWI with CS, they assume that targets are point like positioned at only discrete grid locations and wall thickness and its dielectric constant are perfectly is known. However, in real applications, it is not possible to know the exact target positions or estimate the wall parameters definitely. This work details the theory for CS based TWI and analyzes the imaging ability of CS. In addition, in this work, the effect of errors in unknown parameters on the imaging performance is analyzed and possible solutions are discussed.}, booktitle={2012 20th Signal Processing and Communications Applications Conference, SIU 2012, Proceedings}, author={Duman, M. and Gürbüz, A.C.}, year={2012} }
@inproceedings{tuncer_gurbuz_2011, title={Analysis of unknown velocity and target off the grid problems in compressive sensing based subsurface imaging}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-80051642892&partnerID=MN8TOARS}, DOI={10.1109/ICASSP.2011.5947086}, abstractNote={Sparsity of target space in subsurface imaging problem is used within the framework of the compressive sensing (CS) theory in recent publications to decrease the data acquisition load in practical systems. The developed CS based imaging methods are based on two important assumptions; namely, that the speed of propagation in the medium is known and that potential targets are point like targets positioned at discrete spatial points. However, in most subsurface imaging problems these assumptions are not always valid. The propagation velocity may only be known approximately, and targets will generally not fall on the grid exactly. In this work, the performance of the CS based subsurface imaging methods are analyzed for the above defined problems and possible solutions are discussed.}, booktitle={ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings}, author={Tuncer, M.A.C. and Gurbuz, A.C.}, year={2011}, pages={2880–2883} }
@inproceedings{karaku?_gürbüz_2011, title={Comparison of iterative sparse recovery algorithms,Yi̇nelemeli̇ seyrek geri̇ oluşturma algori̇ tmalarinin karşilaştirilmasi}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-79960385604&partnerID=MN8TOARS}, DOI={10.1109/SIU.2011.5929787}, abstractNote={In this work, we compare two different geometric feature extraction methods derived from coordinates of facial points tracked by Active Appearance Models. The compared feature extraction methods differ in their use of coordinates or distances between facial points and whether they use the information of a neutral facial expression. Experiments on the extended Cohn-Kanade database show that the coordinate-based features using the neutral frame information gives the best emotion recognition results (%94) using a SVC classifier with a polynomial kernel.}, booktitle={2011 IEEE 19th Signal Processing and Communications Applications Conference, SIU 2011}, author={Karaku?, C. and Gürbüz, A.C.}, year={2011}, pages={857–860} }
@article{gurbuz_2012, title={Determination of background distribution for ground-penetrating radar data}, volume={9}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84861201789&partnerID=MN8TOARS}, DOI={10.1109/LGRS.2011.2174137}, abstractNote={Ground-penetrating radars (GPRs) show promising results for subsurface buried target detection. However, the online detection as the GPR scans a region is a difficult problem, and the best performance requires to know the characteristics of the clutter and noise which affect the used test statistics in detection. In statistical detection methods developed for GPR, mostly Gaussian clutter assumption is used mainly due to its simplicity. In this letter, a low-complexity goodness-of-fit test suitable for online GPR detection is applied to experimental GPR data sets to determine the best clutter distribution defining the data test statistic. The distributions of A-scan energies after background subtraction are determined from different experimental data taken over notarget regions. The obtained results show that the GPR clutter for the tested experimental data is mainly gamma distributed than Gaussian. The demonstrated procedure can be applied to any GPR data set for the determination of the background distribution, for target detection, and for selecting detection thresholds properly for GPR applications and more realistic GPR clutter generation simulations.}, number={4}, journal={IEEE Geoscience and Remote Sensing Letters}, author={Gurbuz, A.C.}, year={2012}, pages={544–548} }
@inproceedings{duman_gürbüz_2011, title={Developing data simulation platform and signal processing technics for ground penetrating radar,Yere i̇şleyen radar i̇çi̇n veri benzetim ortamive sinyal i̇şleme yontemlerinin geli̇ştirilmesi̇}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-79960403563&partnerID=MN8TOARS}, DOI={10.1109/SIU.2011.5929620}, abstractNote={Ensemble algorithms have been a very popular research topic because of their high performances. In this work, performance based ensemble pruning and decision weighting methods are investigated on 3 ensemble algorithms (Bagging, Random Subspaces, Random Forest) over 26 classification datasets. According to our experiments; the algorithm including most diversity among its base learners is Random Subspaces. The best performed ensemble algorithm is Random Subspaces with decision weighting.}, booktitle={2011 IEEE 19th Signal Processing and Communications Applications Conference, SIU 2011}, author={Duman, M. and Gürbüz, A.C.}, year={2011}, pages={190–193} }
@article{tuncer_gurbuz_2012, title={Ground reflection removal in compressive sensing ground penetrating radars}, volume={9}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84655163868&partnerID=MN8TOARS}, DOI={10.1109/LGRS.2011.2158981}, abstractNote={Recent results in compressive sensing (CS)-based subsurface imaging showed that, if the target space is sparse, it can be reconstructed with many fewer number of measurements from a stepped frequency ground penetrating radar (GPR). One of the problems in this CS subsurface imaging is the surface reflections. Previous work dealed with surface reflections using a model dictionary generated from the target space excluding specifically the near surface region. While this works fine for some applications, it might lack the imaging of near surface targets. Removing the surface reflections with standard methods is not directly applicable since only very few and random measurements in the frequency domain are taken. This letter provides a simple surface reflection method using compressive measurements, that can be used for nonplanar surfaces. It is observed in both simulated and experimental GPR data that the CS-based imaging method is more robust and can find shallow targets using the surface-reflection-removed data.}, number={1}, journal={IEEE Geoscience and Remote Sensing Letters}, author={Tuncer, M.A.C. and Gurbuz, A.C.}, year={2012}, pages={23–27} }
@article{gürbüz_2011, title={Line detection with adaptive random samples}, volume={19}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-78650876383&partnerID=MN8TOARS}, DOI={10.3906/elk-0910-272}, abstractNote={This paper examines the detection of parameterized shapes in multidimensional noisy grayscale images. A novel shape detection algorithm utilizing random sample theory is presented. Although the method can be generalized, line detection is detailed. Each line in the image corresponds to a point in the line parameter space. The method creates hypothesis lines by randomly selecting parameter space points and tests the surrounding regions for acceptable linear features. The information obtained from each randomly selected line is used to update the parameter distribution, which reducesi the required number of random trials. The selected lines are re-estimated within a smaller search space with a more accurate algorithm like the Hough transform (HT). Faster results are obtained compared to HT, without losing performance as in other faster HT variants. The method is robust and suitable for binary or grayscale images. Results are given from both simulated and experimental subsurface seismic and ground penetrating radar (GPR) images when searching for features like pipes or tunnels.}, number={1}, journal={Turkish Journal of Electrical Engineering and Computer Sciences}, author={Gürbüz, A.C.}, year={2011}, pages={21–32} }
@inproceedings{gürbüz_pilanci_arikan_2011, title={Sparse signal reconstruction with ellipsoid enlargement,Elli̇psoi̇d geni̇şletmeyle seyrek si̇nyal geri̇ oluşturma}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-79960398965&partnerID=MN8TOARS}, DOI={10.1109/SIU.2011.5929771}, abstractNote={Attribute based approach is a new object classification model. The fundamental difference from traditional models is that it employs an attribute layer in the classifier cascade which serves as a switching entity between low level pixel data and high level object labels. The model brings new insights to object classification that we do not observe with the traditional approaches: classification of unseen images, description of the unclassified objects, description of unexpected attributes, description of missing attributes and learning from textual descriptions. Recent preliminary publications give promising results. However, they are not at desired accuracy levels yet. In this work, effects of ternary and probabilistic representations of attributes instead of binary on classification performance are evaluated.}, booktitle={2011 IEEE 19th Signal Processing and Communications Applications Conference, SIU 2011}, author={Gürbüz, A.C. and Pilanci, M. and Arikan, O.}, year={2011}, pages={793–796} }
@inproceedings{tuncer_gürbüz_2011, title={Surface reflection removal in compressed sensing GPR and sparse subsurface imaging,Sikiştirilmiş algilamali GPR i̇çi̇n yüzey yansimalarinin çikarilmasi ve seyrek yeralti görüntüleme}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-79960420970&partnerID=MN8TOARS}, DOI={10.1109/SIU.2011.5929614}, abstractNote={This study focused on the classification of chaotic circuit behaviors with probabilistic neural network (PNN). Although, chaotic circuit outputs track similar traces for the defined parameters, still the circuit outputs preserve their own random characteristics at each trial. PNN is an effective tool for classification of pattern recognition problems. Inherited features of PNN are very compatible with the chaotic circuit output classification problem and it provides satisfying performance. The selection of the proper features in the feature extraction step defines the performance of the classification significantly. In order to, compare classification performance of the PNN, different feature vectors are employed in the training process. Moreover, the spread parameter is a considerably vital factor for the performance of the network. The simulation results and the corresponding illustrations for the performance analysis are also given.}, booktitle={2011 IEEE 19th Signal Processing and Communications Applications Conference, SIU 2011}, author={Tuncer, M.A.C. and Gürbüz, A.C.}, year={2011}, pages={166–169} }
@inproceedings{ayas_gürbüz_2010, title={Analysis of required measurement number in compressive sensing,Sikiştirilmiş algilamada gerekli ölçüm sayisinin analizi}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-78651429328&partnerID=MN8TOARS}, DOI={10.1109/SIU.2010.5653922}, abstractNote={Compressive sensing is a new signal processing method which shows that a sparse signal can be constructed using fewer measurements than normal reconstruction methods. Rather taking all Nyquist samples of a sparse signal in any base the signal can be reconstructed correctly by taking small number of linear projections. In compressive sensing an important relation between measurement number and signal length and sparsity level is used as M=K(logN). This relation is examined separately in noiseless and noisy data. It is determined by simulations that this relation is valid for sparse enough signals and a new relation has been developed for more general cases.}, booktitle={SIU 2010 - IEEE 18th Signal Processing and Communications Applications Conference}, author={Ayas, L. and Gürbüz, A.C.}, year={2010}, pages={914–917} }
@article{compressive sensing of underground structures using gpr_2012, volume={22}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-81555207188&partnerID=MN8TOARS}, DOI={10.1016/j.dsp.2010.11.003}, number={1}, journal={Digital Signal Processing: A Review Journal}, year={2012}, pages={66–73} }
@inproceedings{tuncer_gürbüz_2010, title={Sparsity enhanced fast subsurface imaging for stepped frequency GPRs,Basamak frekansli GPR için seyreklik tabanli hizli yer alti görüntüleme}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-78651462928&partnerID=MN8TOARS}, DOI={10.1109/SIU.2010.5651600}, abstractNote={A sparsity enhanced and fast data acquisition and imaging method is presented for stepped-frequency continuous-wave ground penetrating radars (SFCW GPRs). In previous work it is shown that if the target space is sparse like the point like targets, an image of the target space can be constructed with making measurements at only a small number of random frequencies by solving an l 1 minimization problem. This greatly reduces the data acquisition time but the computational complexity for the imaging method is high. In this work, subsurface imaging is done with a suboptimal but fast method, orthogonal matching pursuit. Similar results to l 1 minimization images are obtained within much shorter times. Also the results are sparse and less cluttered compared to standard backprojection images.}, booktitle={SIU 2010 - IEEE 18th Signal Processing and Communications Applications Conference}, author={Tuncer, M.A.?. and Gürbüz, A.C.}, year={2010}, pages={443–446} }
@inproceedings{gürbüz_2010, title={Sparsity enhanced fast subsurface imaging with GPR}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-77956498731&partnerID=MN8TOARS}, DOI={10.1109/ICGPR.2010.5550130}, abstractNote={Sparsity of a signal starts to become very important in many applications. In subsurface imaging, generally potential targets covers a small part of the total subsurface volume to be imaged, thus the targets are spatially sparse. Under this assumption it is shown that the subsurface imaging problem can be formulated as a dictionary selection problem which can be solved quickly using basis pursuit type algorithms compared to previously published convex optimization based methods. Spatial sparsity also indicates that the number of measurements (spatial or time/frequency) that GPR collects can be reduced, decreasing the data acquisition time. Orthogonal matching pursuit algorithm is used for reconstructing sparse subsurface images. Results show that the proposed method reduces time both in data acquisition and processing compared to previous methods with similar performance.}, booktitle={Proceedings of the 13th Internarional Conference on Ground Penetrating Radar, GPR 2010}, author={Gürbüz, A.C.}, year={2010} }
@article{gurbuz_mcclellan_scott_2009, title={A compressive sensing data acquisition and imaging method for stepped frequency GPRs}, volume={57}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-67650122749&partnerID=MN8TOARS}, DOI={10.1109/TSP.2009.2016270}, abstractNote={ A novel data acquisition and imaging method is presented for stepped-frequency continuous-wave ground penetrating radars (SFCW GPRs). It is shown that if the target space is sparse, i.e., a small number of point like targets, it is enough to make measurements at only a small number of random frequencies to construct an image of the target space by solving a convex optimization problem which enforces sparsity through $\ell _{1}$ minimization. This measurement strategy greatly reduces the data acquisition time at the expense of higher computational costs. Imaging results for both simulated and experimental GPR data exhibit less clutter than the standard migration methods and are robust to noise and random spatial sampling. The images also have increased resolution where closely spaced targets that cannot be resolved by the standard migration methods can be resolved by the proposed method. }, number={7}, journal={IEEE Transactions on Signal Processing}, author={Gurbuz, A.C. and McClellan, J.H. and Scott, W.R.}, year={2009}, pages={2640–2650} }
@article{compressive sensing for subsurface imaging using ground penetrating radar_2009, volume={89}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-67349160257&partnerID=MN8TOARS}, DOI={10.1016/j.sigpro.2009.03.030}, number={10}, journal={Signal Processing}, year={2009}, pages={1959–1972} }
@article{detection of linear and planar structures in 3d subsurface images by iterative dimension reduction_2010, volume={20}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-76349124675&partnerID=MN8TOARS}, DOI={10.1016/j.dsp.2009.06.020}, number={2}, journal={Digital Signal Processing: A Review Journal}, year={2010}, pages={391–400} }
@inproceedings{gurbuz_2009, title={Shape detection in images exploiting sparsity}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-73949120112&partnerID=MN8TOARS}, DOI={10.1109/ISCIS.2009.5291916}, abstractNote={Detection of different kinds of shapes, i.e. lines, circles, hyperbolas etc., in varying kinds of images arises in diverse areas such as signal and image processing, computer vision or remote sensing. The generalized Hough Transform is a traditional approach to detect a specific shape in an image by transforming the problem into a parameter space representation. In this paper we use the observation that the number of shapes in an image is much smaller than the number of all possible shapes. This means the shapes are sparse in the parameter domain. Rather than forming the parameter space from the image as in the HT, we take a reverse approach and ask "which combination of parameter space cells represent my data best?". This leads us to generate a dictionary of shapes and use additional information about sparsity of shapes within a basis pursuit framework. The results indicate enhanced shape detection performance, increased resolution, joint detection of different shapes in an image and robustness to noise. In addition to this, combining the sparsity of shapes with the Compressive Sensing ideas shows that it is possible to directly find the shapes in an image from small number of random projections of the image without first reconstructing the image itself.}, booktitle={2009 24th International Symposium on Computer and Information Sciences, ISCIS 2009}, author={Gurbuz, A.C.}, year={2009}, pages={70–75} }
@inproceedings{gürbüz_mcclellan_cevher_2008, title={A compressive beamforming method}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-51449093488&partnerID=MN8TOARS}, DOI={10.1109/ICASSP.2008.4518185}, abstractNote={Compressive Sensing (CS) is an emerging area which uses a relatively small number of non-traditional samples in the form of randomized projections to reconstruct sparse or compressible signals. This paper considers the direction-of-arrival (DOA) estimation problem with an array of sensors using CS. We show that by using random projections of the sensor data, along with a full waveform recording on one reference sensor, a sparse angle space scenario can be reconstructed, giving the number of sources and their DOA's. The number of projections can be very small, proportional to the number sources. We provide simulations to demonstrate the performance and the advantages of our compressive beamformer algorithm.}, booktitle={ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings}, author={Gürbüz, A.C. and McClellan, J.H. and Cevher, V.}, year={2008}, pages={2617–2620} }
@inproceedings{compressive sensing of parameterized shapes in images_2008, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-51549084025&partnerID=MN8TOARS}, DOI={10.1109/ICASSP.2008.4518018}, abstractNote={Compressive Sensing (CS) uses a relatively small number of non-traditional samples in the form of randomized projections to reconstruct sparse or compressible signals. The Hough transform is often used to find lines and other parameterized shapes in images. This paper shows how CS can be used to find parameterized shapes in images, by exploiting sparseness in the Hough transform domain. The utility of the CS-based method is demonstrated for finding lines and circles in noisy images, and then examples of processing GPR and seismic data for tunnel detection are presented.}, booktitle={ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings}, year={2008}, pages={1949–1952} }
@inproceedings{cevher_gurbuz_mcclellan_chellappa_2008, title={Compressive wireless arrays for bearing estimation}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-51449109098&partnerID=MN8TOARS}, DOI={10.1109/ICASSP.2008.4518155}, abstractNote={Joint processing of sensor array outputs improves the performance of parameter estimation and hypothesis testing problems beyond the sum of the individual sensor processing results. When the sensors have high data sampling rates, arrays are tethered, creating a disadvantage for their deployment and also limiting their aperture size. In this paper, we develop the signal processing algorithms for randomly deployable wireless sensor arrays that are severely constrained in communication bandwidth. We focus on the acoustic bearing estimation problem and show that when the target bearings are modeled as a sparse vector in the angle space, low dimensional random projections of the microphone signals can be used to determine multiple source bearings by solving an ℓ 1 -norm minimization problem. Field data results are shown where only 10 bits of information is passed from each microphone to estimate multiple target bearings.}, booktitle={ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings}, author={Cevher, V. and Gurbuz, A.C. and McClellan, J.H. and Chellappa, R.}, year={2008}, pages={2497–2500} }
@inproceedings{gpr imaging using compressed measurements_2008, volume={2}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-66549129325&partnerID=MN8TOARS}, DOI={10.1109/IGARSS.2008.4778915}, abstractNote={A new data acquisition and imaging method exploiting the sparsity of the target space is presented for ground penetrating radar (GPR) imaging. Sparsity is enforced by solving a convex l 1 minimization problem which uses a very small number of random measurements. The method can greatly reduce the data acquisition time while producing sparse target space images. Simulation and experimental data results are provided to show that the method has excellent resolution and is robust to noise and random spatial sampling.}, number={1}, booktitle={International Geoscience and Remote Sensing Symposium (IGARSS)}, year={2008} }
@inproceedings{compressive sensing for gpr imaging_2007, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-50249177603&partnerID=MN8TOARS}, DOI={10.1109/ACSSC.2007.4487636}, abstractNote={The theory of compressive sensing (CS) enables the reconstruction of sparse signals from a small set of non-adaptive linear measurements by solving a convex l 1 minimization problem. This paper presents a novel data acquisition and imaging algorithm for Ground Penetrating Radars (GPR) based on CS by exploiting sparseness in the target space, i.e., a small number of point-like targets. Instead of measuring conventional radar returns and sampling at the Nyquist rate, linear projections of the returned signal with random vectors are taken as measurements. Using simulated and experimental GPR data, it is shown that sparser and sharper target space images can be obtained compared to standard backprojection methods using only a small number of CS measurements. Furthermore, the target region can even be sampled at random aperture points.}, booktitle={Conference Record - Asilomar Conference on Signals, Systems and Computers}, year={2007}, pages={2223–2227} }
@inproceedings{gurbuz_mcclellan_scott_2007, title={Detecting curved underground tunnels using partial radon transforms}, volume={1}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-34547540286&partnerID=MN8TOARS}, DOI={10.1109/ICASSP.2007.365965}, abstractNote={The Radon Transform (RT) is known to be effective in detecting lines in noisy images, but it is not capable of detecting curves unless the curve parametrization is given. In this paper, partial Radon transforms (PRT) are investigated as a tool to detect curved features such as underground tunnels in ground penetrating radar (GPR) images. The algorithm applies the Radon Transform to small batches of the total image and updates the tunnel position parameters as new batches are used. Missing data, as well as finding the ends of tunnels can be handled with the proposed algorithm. Performance analysis is given for various signal-to-noise ratios (SNR) and batch sizes. The effect of the curvature level on the performance is also analyzed.}, booktitle={ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings}, author={Gurbuz, A.C. and McClellan, J.H. and Scott, W.R.}, year={2007} }
@inproceedings{gürbüz_mcclellan_scott_2007, title={Detecting features using random sample theory,Rastgele örnek teorisi kullanilarak şeki̇l bulma}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-50249138256&partnerID=MN8TOARS}, DOI={10.1109/SIU.2007.4298779}, abstractNote={This paper aims to detect features in 2-D and 3-D highly noisy images using random sample theory fast and with high detection performance. The proposed method yields faster results than standard feature detection algorithms, such as the Hough Transform (HT) or its variants, while keeping the the performance level of HT. Proposed method first finds possible feature areas by creating random hypothesis and testing them. Features are re-estimated by only searching these possible areas which reduces the total search space. The proposed algorithm is tested on both simulated and experimental subsurface Seismic and GPR images for searching linear features like pipes or tunnels. Results show that the proposed algorithm can detect features accurately and much faster than conventional methods.}, booktitle={2007 IEEE 15th Signal Processing and Communications Applications, SIU}, author={Gürbüz, A.C. and McClellan, J.H. and Scott, W.R.}, year={2007} }
@inproceedings{feature detection in highly noisy images using random sample theory_2007, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-47649129017&partnerID=MN8TOARS}, DOI={10.1109/ICDSP.2007.4288609}, abstractNote={A novel feature detection algorithm utilizing random sample theory is proposed for 2D and 3D images. The proposed method works on both binary and gray-scale images and yields faster results than standard feature detection algorithms, such as the Hough Transform (HT), while keeping the performance level of HT. The proposed method creates random hypothesis features and tests them to select candidate features in the image. The selected candidate features are then re-estimated within a smaller search space around the candidate feature. The proposed algorithm has been tested on both simulated and experimental subsurface seismic and GPR images to locate linear features like pipes or tunnels. Results show that the proposed algorithm can detect features accurately and much faster than conventional methods.}, booktitle={2007 15th International Conference on Digital Signal Processing, DSP 2007}, year={2007}, pages={423–426} }
@inproceedings{feature detection in images by adaptive random sampling_2007, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-47849099024&partnerID=MN8TOARS}, DOI={10.1109/SSP.2007.4301327}, abstractNote={Random sample theory is an effective tool for detecting features in images. This paper presents an adaptive random sampling scheme that clusters random samples into candidate features. The required trial number is reduced by adaptive sampling, thereby reducing the run time of the algorithm. The proposed method quickly finds rough regions in the image that may include features using adaptive random sampling and re-estimates the features using the Hough Transform (HT) within the smaller regions. The proposed algorithm is tested on both simulated and experimental subsurface seismic and GPR images to search for linear features like pipes or tunnels. Faster results are obtained as compared to standard feature detection algorithms, such as the HT or its variants, while maintaining the similar performance level as the HT.}, booktitle={IEEE Workshop on Statistical Signal Processing Proceedings}, year={2007}, pages={591–595} }
@inproceedings{investigation of the detection of shallow tunnels using electromagnetic and seismic waves_2007, volume={6553}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-35948976157&partnerID=MN8TOARS}, DOI={10.1117/12.722437}, abstractNote={Multimodal detection of subsurface targets such as tunnels, pipes, reinforcement bars, and structures has been investigated using both ground-penetrating radar (GPR) and seismic sensors with signal processing techniques to enhance localization capabilities. Both systems have been tested in bi-static configurations but the GPR has been expanded to a multi-static configuration for improved performance. The use of two compatible sensors that sense different phenomena (GPR detects changes in electrical properties while the seismic system measures mechanical properties) increases the overall system's effectiveness in a wider range of soils and conditions. Two experimental scenarios have been investigated in a laboratory model with nearly homogeneous sand. Images formed from the raw data have been enhanced using beamforming inversion techniques and Hough Transform techniques to specifically address the detection of linear targets. The processed data clearly indicate the locations of the buried targets of various sizes at a range of depths.}, booktitle={Proceedings of SPIE - The International Society for Optical Engineering}, year={2007} }
@article{multistatic ground-penetrating radar experiments_2007, volume={45}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-34547436721&partnerID=MN8TOARS}, DOI={10.1109/TGRS.2007.900677}, abstractNote={ A multistatic ground-penetrating radar (GPR) system has been developed and used to measure the response of a number of targets to produce data for the investigation of multistatic inversion algorithms. The system consists of a linear array of resistive-vee antennas, microwave switches, a vector network analyzer, and a 3-D positioner, all under computer control. The array has two transmitters and four receivers which provide eight bistatic spacings from 12 to 96 cm in 12-cm increments. Buried targets are scanned with and without surface clutter, which is a layer of rocks whose spacing is empirically chosen to maximize the clutter effect. The measured responses are calibrated so that the direct coupling in the system is removed, and the signal reference point is located at the antenna drive point. Images are formed using a frequency-domain beamforming algorithm that compensates for the phase response of the antennas. Images of targets in air validate the system calibration and the imaging algorithm. Bistatic and multistatic images for the buried targets are very good, and they show the effectiveness of the system and processing. }, number={8}, journal={IEEE Transactions on Geoscience and Remote Sensing}, year={2007}, pages={2544–2553} }
@inproceedings{application of multi-static inversion algorithms to landmine detection_2006, volume={6217 II}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-33747373789&partnerID=MN8TOARS}, DOI={10.1117/12.666367}, abstractNote={Multi-static ground-penetrating radar (GPR) uses an array of antennas to conduct a number of bistatic operations simultaneously. The multi-static GPR is used to obtain more information on the target of interest using angular diversity. An entirely computer controlled, multi-static GPR consisting of a linear array of six resistively-loaded vee dipoles (RVDs), a network analyzer, and a microwave switch matrix was developed to investigate the potential of multi-static inversion algorithms. The performance of a multi-static inversion algorithm is evaluated for targets buried in clean sand, targets buried under the ground covered by rocks, and targets held above the ground (in the air) using styrofoam supports. A synthetic-aperture, multi-static, time-domain GPR imaging algorithm is extended from conventional mono-static back-projection techniques and used to process the data. Good results are obtained for the clean surface and air targets; however, for targets buried under rocks, only the deeply buried targets could be accurately detected and located.}, booktitle={Proceedings of SPIE - The International Society for Optical Engineering}, year={2006} }
@inproceedings{combined ground penetrating radar and seismic system for detecting tunnels_2006, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-34547464985&partnerID=MN8TOARS}, DOI={10.1109/IGARSS.2006.318}, abstractNote={An experimental system to collect co-located ground penetrating radar (GPR) and seismic data was developed to investigate possibilities of using the sensors individually or in a cooperative manner to detect shallow tunnels. These sensors were chosen because they sense very different physical properties. The seismic sensor is sensitive to the differences between the mechanical properties of a tunnel and the soil while the GPR is sensitive to the dielectric properties. Raw and processed data from both sensors are presented.}, booktitle={International Geoscience and Remote Sensing Symposium (IGARSS)}, year={2006}, pages={1232–1235} }
@inproceedings{imaging of subsurface targets using a 3d quadtree algorithm_2005, volume={IV}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-33646759969&partnerID=MN8TOARS}, DOI={10.1109/ICASSP.2005.1416206}, abstractNote={The imaging of subsurface targets using ground penetrating radar (GPR) is becoming an increasingly important area of research. Conventional image formation techniques expend large amounts of computation to resolve a region fully, even a region of clutter. However, by using multi-resolution techniques, e.g, quadtree algorithms, potential targets and clutter can be discriminated in a computationally efficient way. Prior work has focused on the development of 2D quadtree algorithms for surface targets. For mine detection, target depth adds another dimension; thus, we have developed a 3D quadtree algorithm, and applied a multi-stage detector that uses the energy change between quadtree stages to discriminate target and clutter regions. This algorithm is then tested on computer-generated data, as well as experimental data collected from a model mine field. Results show that target location information can be obtained even under near field and small aperture conditions.}, booktitle={ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings}, year={2005} }
@inproceedings{imaging of subsurface targets using a 3d quadtree algorithm,yeralti hedeflerinin 3d dörtlü aǧaç algoritmasi ile görüntülenmesi_2005, volume={2005}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-33846611994&partnerID=MN8TOARS}, DOI={10.1109/SIU.2005.1567742}, booktitle={Proceedings of the IEEE 13th Signal Processing and Communications Applications Conference, SIU 2005}, year={2005}, pages={544–547} }
@inproceedings{gurbuz_mcclellan_2006, title={Iterative detection of linear objects in GPR and seismic images}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-34250721490&partnerID=MN8TOARS}, DOI={10.1109/SAM.2006.1677170}, booktitle={2006 IEEE Sensor Array and Multichannel Signal Processing Workshop Proceedings, SAM 2006}, author={Gurbuz, A.C. and McClellan, J.H.}, year={2006}, pages={118–121} }
@inproceedings{gürbüz_mcclellan_scott_2006, title={Locating subsurface targets using minimal GPR measurements,Yeralti hedefleri̇ni̇n asgari̇ GPR ölçü mleri̇ i̇le tesbi̇ti̇}, volume={2006}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-34247095136&partnerID=MN8TOARS}, DOI={10.1109/SIU.2006.1659746}, abstractNote={This paper describes a novel approach which finds probable target areas in ground penetrating radars (GPR) with minimal number of measurements. A subsurface GPR target creates a hyperbolic signature in the space-time domain. Conventional algorithms use these signatures to image and detect subsurface targets. Our approach uses the time delay differences (TDD) in consecutive GPR A-Scan measurements to localize the target. For further reduction of measurement numbers regions of low target probability are determined to eliminate redundant measurements. Results from the experimental data from a model mine field show that the target positions can be found accurately using much fewer measurements than the conventional imaging algorithms.}, booktitle={2006 IEEE 14th Signal Processing and Communications Applications Conference}, author={Gürbüz, A.C. and McClellan, J.H. and Scott, W.R.}, year={2006} }
@inproceedings{predicting gpr target locations using time delay differences_2006, volume={6217 II}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-33747341831&partnerID=MN8TOARS}, DOI={10.1117/12.666335}, abstractNote={We describe an efficient approach for finding probable target areas quickly with a minimal number of Ground Penetrating Radar (GPR) measurements. Since a potential GPR target creates a hyperbolic signature in the space-time domain, our approach uses the time delay differences from consecutive GPR A-Scan data to estimate the location of the apex of the hyperbolic signature, thus locating a target. This apex prediction method uses many fewer measurements than a full backprojection algorithm. Regions of low target probability are determined using a Neyman-Pearson detection approach in order to eliminate redundant measurements. In this regard, our approach is especially suitable as a pre-screener: other sensors that are more accurate, but require more measurement time, can then be applied only to high probability-of-target areas to corroborate results, differentiate between targets, or provide more accurate location measurements. Compared to a standard backprojection algorithm more signal-to-noise ratio (SNR) is needed to achieve similar detection performance. This SNR loss can be reduced by using a more conservative algorithm which reduces the step size of the GPR antenna. Results from experimental data collected at a model mine field at the Georgia Institute of Technology show that target positions can be found accurately using less than 10% of the measurements utilized by conventional imaging algorithms.}, booktitle={Proceedings of SPIE - The International Society for Optical Engineering}, year={2006} }
@inproceedings{gürbüz_mcclellan_scott_larson_2006, title={Seismic imaging and detection of underground tunnels,Si̇smi̇k tünel görüntülenmesi̇ ve algilanmasi}, volume={2006}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-34247140580&partnerID=MN8TOARS}, DOI={10.1109/SIU.2006.1659837}, abstractNote={To investigate the problem of detecting and imaging underground tunnels, an experimental system that utilizes seismic waves has been constructed. Seismic reflections from the tunnel are transformed into a 3D image using a synthetic aperture time-delay backprojection algorithm. Results from experimental data show that the tunnel is directly visible in the backprojected image. Nevertheless, tunnels with low signal to noise ratio (SNR) are located using 2D and 3D Radon Transforms followed by a detection algorithm. A simulation is performed on the performance of the Radon transform for detecting lines in noisy images and it is shown how lines in very low SNR images can be detected. Also it is observed that longer lines have higher probability of detection at the same noise level.}, booktitle={2006 IEEE 14th Signal Processing and Communications Applications Conference}, author={Gürbüz, A.C. and McClellan, J.H. and Scott, W.R. and Larson, G.D.}, year={2006} }
@inproceedings{seismic tunnel imaging and detection_2006, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-50249106646&partnerID=MN8TOARS}, DOI={10.1109/ICIP.2006.312911}, abstractNote={To investigate the problem of detecting and imaging underground tunnels, an experimental system that utilizes seismic waves has been constructed. Seismic reflections from the tunnel are transformed into a 3D image using a synthetic aperture time-delay backprojection algorithm. Results from experimental data show that the tunnel is directly visible in the backprojected image. Nevertheless, tunnels with low signal to noise ratio (SNR) are located using 2D and 3D Radon Transforms followed by a detection algorithm. A simulation is performed on the performance of the Radon transform for detecting lines in noisy images and it is shown how lines in very low SNR images can be detected. Also it is observed that longer lines have higher probability of detection at the same noise level.}, booktitle={Proceedings - International Conference on Image Processing, ICIP}, year={2006}, pages={3229–3232} }
@inproceedings{a multi-static ground-penetrating radar with an array of resistively-loaded vee dipole antennas for landmine detection_2005, volume={5794}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-26844516285&partnerID=MN8TOARS}, DOI={10.1117/12.604145}, abstractNote={A multi-static ground-penetrating radar (GPR) has been developed to investigate the potential of multi-static inversion algorithms. The GPR consists of a linear array of six resistively-loaded vee dipoles (RVDs), a network analyzer, and a microwave switch matrix all under computer control. The antennas in the array are spaced 12cm apart so the spacing between the transmitter and the receiver pairs in the measurements are from 12cm to 96cm in 12cm increments. The size of the array is suitable for the landmine problem and scaled measurements of the buried structure problem. The RVD is chosen as an array element because it is very "clean" in that it has very little self clutter and a very low radar cross section to lessen the reflections between the ground and the antenna. The shape and the loading profile of the antenna are designed to decrease the reflection at the drive point of the antenna while increasing the forward gain. The antenna and balun are made in a module, which is mechanically reliable without significant performance degradation. The multi-static GPR operation is demonstrated on targets buried in clean sand and targets buried under the ground covered by rocks. The responses of the targets are measured by each transmitter-receiver pair. A synthetic aperture, multi-static GPR imaging algorithm is extended from conventional monostatic back-projection techniques and used to process the data. Initial images obtained from the multi-static data are clearer than those obtained from bistatic data.}, number={PART I}, booktitle={Proceedings of SPIE - The International Society for Optical Engineering}, year={2005}, pages={495–506} }
@inproceedings{subsurface target imaging using a multi-resolution 3d quadtree algorithm_2005, volume={5794}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-26844532539&partnerID=MN8TOARS}, DOI={10.1117/12.603748}, abstractNote={The imaging of subsurface targets, such as landmines, using Ground Penetrating Radar (GPR) is becoming an increasingly important area of research. Conventional image formation techniques expend large amounts of computational resources on fully resolving a region, even if there is a large amount of clutter. For example, standard backprojection algorithms require O(N3). However, by using multi-resolution techniques-such as quadtree-potential targets and clutter can be discriminated more efficiently with O(N2log2N). Because prior work has focused on the imaging of surface targets, quadtree techniques have mostly been developed for 2D imaging. Target depth adds another dimension to the imaging problem; therefore, we have developed a 3D quadtree algorithm. In this case, the mine field is modeled as a volume that is sub-divided at each stage of the quadtree algorithm. From each of these sub-volumes, the energy intensity is calculated. As the algorithm proceeds to finer resolutions, the energy in region containing a potential target increases, while that of background noise decreases. A multi-stage detector applied on intermediate quadtree data uses this change in energy to discriminate between regions of targets and clutter. This is advantageous because only the regions containing likely targets are investigated by additional sensors that are relatively slow in comparison to GPR (e.g. seismic or EMI sensors). This algorithm is tested on synthetic and experimental data collected from a model mine field at Georgia Institute of Technology. Even under near field and small aperture conditions, which hold for the mine detection case, test results show that target location information can be gathered with processing using the 3D quadtree algorithm.}, number={PART II}, booktitle={Proceedings of SPIE - The International Society for Optical Engineering}, year={2005}, pages={1172–1181} }
@inproceedings{combined seismic, radar, and induction sensor for landmine detection_2004, volume={3}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-15944422479&partnerID=MN8TOARS}, booktitle={International Geoscience and Remote Sensing Symposium (IGARSS)}, year={2004}, pages={1613–1616} }