@article{ozturk_anjinappa_erden_chowdhury_guvenc_dai_bhuyan_2023, title={Channel Rank Improvement in Urban Drone Corridors Using Passive Intelligent Reflectors}, ISSN={["1095-323X"]}, DOI={10.1109/AERO55745.2023.10115741}, abstractNote={Multiple-input multiple-output (MIMO) techniques can help in scaling the achievable air-to-ground (A2G) channel capacity while communicating with drones. However, spatial multiplexing with drones suffers from rank-deficient channels due to the unobstructed line-of-sight (LoS), especially in millimeter-wave (mmWave) frequencies that use narrow beams. One possible solution is utilizing low-cost and low-complexity metamaterial-based intelligent reflecting surfaces (IRS) to enrich the multi path environment, taking into account that the drones are restricted to flying only within well-defined drone corridors. A hurdle with this solution is placing the IRSs optimally. In this study, we propose an approach for IRS placement with a goal to improve the spatial multiplexing gains, and hence, to maximize the average channel capacity in a predefined drone corridor. Our results at 6 GHz, 28 GHz, and 60 GHz show that the proposed approach increases the average rates for all frequency bands for a given drone corridor when compared with the environment with no IRSs present, and IRS-aided channels perform close to each other at sub-6 and mmWave bands.}, journal={2023 IEEE AEROSPACE CONFERENCE}, author={Ozturk, Ender and Anjinappa, Chethan K. and Erden, Fatih and Chowdhury, Md Moin Uddin and Guvenc, Ismail and Dai, Huaiyu and Bhuyan, Arupjyoti}, year={2023} }
 @article{ozturk_erden_du_anjinappa_ozdemir_guvenc_2022, title={Ray Tracing Analysis of Sub-6 GHz and mmWave Indoor Coverage with Reflecting Surfaces}, ISSN={["2164-2958"]}, DOI={10.1109/RWS53089.2022.9719917}, abstractNote={Indoor coverage and channel modelling is crucial for network planning purposes at mmWave bands. In this paper, we analyzed received power patterns and connectivity in an indoor office environment for sub-6 GHz and mmWave bands using ray tracing simulations and theoretical models over different scenarios. We discussed the effect of using metallic walls instead of regular drywall, base station (BS) location, and open/shut doors. Our results showed that ray tracing solutions are consistent with theoretical calculations, and using reflective walls significantly improves average received power and connectivity at mmWave bands, e.g., for the given floor plan, coverage increases from 86% to 97.5% at 60 GHz band.}, journal={2022 IEEE RADIO AND WIRELESS SYMPOSIUM (RWS)}, author={Ozturk, Ender and Erden, Fatih and Du, Kairui and Anjinappa, Chethan K. and Ozdemir, Ozgur and Guvenc, Ismail}, year={2022}, pages={160–163} }
 @article{du_ozdemir_erden_guvenc_2021, title={28 GHz Indoor and Outdoor Propagation Measurements and Analysis at a Regional Airport}, DOI={10.1109/PIMRC50174.2021.9569260}, abstractNote={In the upcoming 5G communication, the millimeter-wave (mmWave) technology will play an important role due to its large bandwidth and high data rate. However, mmWave frequencies have higher free-space path loss (FSPL) in line-of-sight (LOS) propagation compared to the currently used sub-6 GHz frequencies. What is more, in non-line-of-sight (NLOS) propagation, the attenuation of mmWave is larger compared to the lower frequencies, which can seriously degrade the performance. It is therefore necessary to investigate mmWave propagation characteristics for different deployment scenarios of interest, to understand coverage and rate performance in such scenarios. In this paper, we focus on 28 GHz wideband mmWave signal propagation characteristics at Johnston Regional Airport (JNX), a local airport near Raleigh, NC. To collect data, we use an NI PXI-based channel sounder at 28 GHz for indoor, outdoor, and indoor-to-outdoor scenarios. Results on LOS propagation, reflection, penetration, signal coverage, and multipath components (MPCs) show a lower indoor FSPL, a richer scattering, and a better coverage compared to outdoor. We also observe high indoor-to-outdoor propagation losses.}, journal={2021 IEEE 32ND ANNUAL INTERNATIONAL SYMPOSIUM ON PERSONAL, INDOOR AND MOBILE RADIO COMMUNICATIONS (PIMRC)}, author={Du, Kairui and Ozdemir, Ozgur and Erden, Fatih and Guvenc, Ismail}, year={2021} }
 @article{anjinappa_erden_guvenc_2021, title={Base Station and Passive Reflectors Placement for Urban mmWave Networks}, volume={70}, ISSN={["1939-9359"]}, url={https://doi.org/10.1109/TVT.2021.3065221}, DOI={10.1109/TVT.2021.3065221}, abstractNote={The use of millimeter-wave (mmWave) bands in 5G networks introduces a new set of challenges to network planning. Vulnerability to blockages and high path loss at mmWave frequencies require careful planning of the network to achieve a desired service quality. In this paper, we propose a novel 3D geometry-based framework for deploying mmWave base stations (gNBs) in urban environments by considering first-order reflection effects. We also provide a solution for the optimum deployment of passive metallic reflectors (PMRs) to extend radio coverage to non-line-of-sight (NLoS) areas. In particular, we perform visibility analysis to find the direct and indirect visibility regions, and using these, we derive a geometry-and-blockage-aided path loss model. We then formulate the network planning problem as two independent optimization problems, placement of gNB(s) and PMRs, to maximize the coverage area, minimize the deployment cost, and maintain a desired quality-of-service level. We test the efficacy of our proposed approach using a generic map and compare our simulation results with the ray tracing solution. Our simulation results show that considering the first-order reflections in planning the mmWave network helps reduce the number of PMRs required to cover the NLoS area. Moreover, the gNB placement aided with PMRs require fewer gNBs to cover the same area, which in turn reduces the deployment cost.}, number={4}, journal={IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Anjinappa, Chethan Kumar and Erden, Fatih and Guvenc, Ismail}, year={2021}, month={Apr}, pages={3525–3539} }
 @article{sennik_erden_constantino_oh_dean_oralkan_2021, title={Electronic nose system based on a functionalized capacitive micromachined ultrasonic transducer (CMUT) array for selective detection of plant volatiles}, volume={341}, ISSN={["0925-4005"]}, url={https://doi.org/10.1016/j.snb.2021.130001}, DOI={10.1016/j.snb.2021.130001}, abstractNote={Here, a small, low-power, wireless gas sensor platform for selective detection of volatile organic compounds (VOCs) released from plants under different abiotic or biotic stress conditions is described. This sensor platform is implemented based on a capacitive micromachined ultrasonic transducer (CMUT) array, in which elements were functionalized with a variety of materials including polymers, phthalocyanines, and metals to improve selectivity. Input impedance measurements of the functionalized CMUT array were compared to pre-coating measurements to analyze the mechanical loading. The CMUT arrays were then exposed to VOCs known to be emitted by plants with different concentrations under dry air flow at room temperature. The results demonstrated that 1-Octanol created the strongest response across different channels and a resolution of 3-ppb was calculated for the CMUT element functionalized using silver ink when exposed to 1-Octanol. The relative responses of different channels to tested volatiles were observed to be different. The k-nearest neighbor (k-NN) algorithm was used for the gas classification by dividing the data to training and test groups. The k-NN results showed that the gases at low concentrations were successfully classified with better than 97 % accuracy. Finally, to emulate the ambient atmosphere for plants, the gas tests were repeated by adding different levels of humidity to the gas flow. With a minimum 98 % accuracy, the k-NN classifier demonstrated that the functionalized CMUT array can be used for selective detection of the group of plant VOCs used in this study, even at different relative humidity levels in the ambient atmosphere.}, journal={SENSORS AND ACTUATORS B-CHEMICAL}, publisher={Elsevier BV}, author={Sennik, Erdem and Erden, Fatih and Constantino, Nasie and Oh, YeonYee and Dean, Ralph A. and Oralkan, Omer}, year={2021}, month={Aug} }
 @article{du_ozdemir_erden_guvenc_2021, title={Sub-Terahertz and mmWave Penetration Loss Measurements for Indoor Environments}, ISSN={["2164-7038"]}, DOI={10.1109/ICCWorkshops50388.2021.9473898}, abstractNote={Millimeter-wave (mmWave) and terahertz (THz) spectrum can support significantly higher data rates compared to lower frequency bands and hence are being actively considered for 5G wireless networks and beyond. These bands have high free-space path loss (FSPL) in line-of-sight (LOS) propagation due to their shorter wavelength. Moreover, in non-line-of-sight (NLOS) scenario, these two bands suffer higher penetration loss than lower frequency bands which could seriously affect the network coverage. It is therefore critical to study the NLOS penetration loss introduced by different building materials at mmWave and THz bands, to help establish link budgets for an accurate performance analysis in indoor environments. In this work, we measured the penetration loss and the attenuation of several common constructional materials at mmWave (28 and 39 GHz) and sub-THz (120 and 144 GHz) bands. Measurements were conducted using a channel sounder based on NI PXI platforms. Results show that the penetration loss changes extensively based on the frequency and the material properties, ranging from 0.401 dB for ceiling tile at 28 GHz, to 16.608 dB for plywood at 144 GHz. Ceiling tile has the lowest measured attenuation at 28 GHz, while clear glass has the highest attenuation of 27.633 dB/cm at 144 GHz. As expected, the penetration loss and attenuation increased with frequency for all the tested materials.}, journal={2021 IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS WORKSHOPS (ICC WORKSHOPS)}, author={Du, Kairui and Ozdemir, Ozgur and Erden, Fatih and Guvenc, Ismail}, year={2021} }
 @article{erden_ozdemir_guvenc_2020, title={28 GHz mmWave Channel Measurements and Modeling in a Library Environment}, volume={2020-January}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85083202769&partnerID=MN8TOARS}, DOI={10.1109/rws45077.2020.9050106}, abstractNote={To fully exploit the millimeter-wave bands for the fifth generation cellular systems, an accurate understanding of the channel propagation characteristics is required, and hence extensive measurement campaigns in different environments are needed. In this paper, we use a rotated directional antenna-based channel sounder for measurements at 28 GHz in large indoor environments at a library setting. We present models for power angular-delay profile and large-scale path loss based on the measurements over distances ranging from 10 m to 50 m. In total, nineteen different line-of-sight (LOS) and non-line-of-sight (NLOS) scenarios are considered, including the cases where the transmitter and the receiver are placed on different floors. Results show that the close-in free space reference distance and the floating intercept path loss models both perform well in fitting the empirical data. The path loss exponent obtained for the LOS scenarios is found to be very close to that of the free space path loss model.}, journal={2020 IEEE Radio and Wireless Symposium (RWS)}, publisher={IEEE}, author={Erden, Fatih and Ozdemir, Ozgur and Guvenc, Ismail}, year={2020}, month={Jan}, pages={52–55} }
 @article{sayeed_vouras_gentile_weiss_quimby_cheng_modad_zhang_anjinappa_erden_et al._2020, title={A Framework for Developing Algorithms for Estimating Propagation Parameters from Measurements}, ISSN={["2166-0069"]}, DOI={10.1109/GCWkshps50303.2020.9367404}, abstractNote={A framework is proposed for developing and evaluating algorithms for extracting multipath propagation components (MPCs) from measurements collected by sounders at millimeter-wave (mmW) frequencies. To focus on algorithmic performance, an idealized model is proposed for the spatial frequency response of the propagation environment measured by a sounder. The input to the sounder model is a pre-determined set of MPC parameters that serve as the “ground truth”. A three-dimensional angle-delay (beamspace) representation of the measured spatial frequency response serves as a natural domain for implementing and analyzing MPC extraction algorithms. Metrics for quantifying the error in estimated MPC parameters are introduced. Initial results are presented for a greedy matching pursuit algorithm that performs a least-squares (LS) reconstruction of the MPC path gains within the iterations. The results indicate that the simple greedy-LS algorithm has the ability to extract MPCs over a large dynamic range, and suggest several avenues for further performance improvement through extensions of the greedy-LS algorithm as well as by incorporating features of other algorithms, such as SAGE and RIMAX.}, journal={2020 IEEE GLOBECOM WORKSHOPS (GC WKSHPS)}, author={Sayeed, Akbar and Vouras, Peter and Gentile, Camillo and Weiss, Alec and Quimby, Jeanne and Cheng, Zihang and Modad, Bassel and Zhang, Yuning and Anjinappa, Chethan and Erden, Fatih and et al.}, year={2020} }
 @article{erden_ozdemir_khawaja_guvenc_2020, title={Correction of Channel Sounding Clock Drift and Antenna Rotation Effects for mmWave Angular Profile Measurements}, volume={1}, ISSN={["2637-6431"]}, url={https://doi.org/10.1109/OJAP.2020.2979243}, DOI={10.1109/OJAP.2020.2979243}, abstractNote={Proper characterization of the millimeter-wave (mmWave) propagation channel requires measuring the power angular-delay profile of the channel which includes angle-of-departure and angle-of-arrival of the multipath components (MPCs). In this paper, we first describe in detail our rotating directional antennas-based 28 GHz channel sounder. Then, for this specific sounder class, we describe and address the following two problems in extracting the MPCs from the measurements: 1) For long-distance channel measurements, triggering signal cannot be generated for the TX and the RX using a single clock (SICL). This necessitates the use of separate clocks (SECLs) which introduces a random timing drift between the clocks. 2) As positions of the antennas change during scanning, total distance traveled by the same MPC differs at each measurement. These problems together cause missing some of the MPCs and detecting MPCs that do not exist in reality. We propose an algorithm to correct the clock drift and MPC delay errors due to the rotation of the antennas. We compare the MPCs from the SICL measurement and the corrected SECL measurements using a Hungarian algorithm based MPC matching method. We show that the percentage of the matched MPCs increases from 28.36% to 74.13% after the correction process.}, journal={IEEE OPEN JOURNAL OF ANTENNAS AND PROPAGATION}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Erden, Fatih and Ozdemir, Ozgur and Khawaja, Wahab and Guvenc, Ismail}, year={2020}, pages={71–87} }
 @article{khawaja_ozdemir_yapici_erden_guvenc_2020, title={Coverage Enhancement for NLOS mmWave Links Using Passive Reflectors}, volume={1}, ISSN={["2644-125X"]}, url={https://doi.org/10.1109/OJCOMS.2020.2969751}, DOI={10.1109/OJCOMS.2020.2969751}, abstractNote={The future 5G networks are expected to use millimeter wave (mmWave) frequency bands to take advantage of the large unused spectrum. However, due to the high path loss at mmWave frequencies, coverage of mmWave signals can get severely reduced, especially for non-line-of-sight (NLOS) scenarios as mmWave signals are severely attenuated when going through obstructions. In this work, we study the use of passive metallic reflectors of different shapes/sizes to improve 28 GHz mmWave signal coverage for both indoor and outdoor NLOS scenarios. We quantify the gains that can be achieved in the link quality with metallic reflectors using measurements, analytical expressions, and ray tracing simulations. In particular, we provide an analytical model for the end-to-end received power in an NLOS scenario using reflectors of different shapes and sizes. For a given size of the flat metallic sheet reflector approaching to the size of the incident beam, we show that the reflected received power for the NLOS link is the same as line-of-sight (LOS) free space received power of the same link distance. Extensive results are provided to study the impact of environmental features and reflector characteristics on NLOS link quality.}, journal={IEEE OPEN JOURNAL OF THE COMMUNICATIONS SOCIETY}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Khawaja, Wahab and Ozdemir, Ozgur and Yapici, Yavuz and Erden, Fatih and Guvenc, Ismail}, year={2020}, pages={263–281} }
 @article{ezuma_erden_anjinappa_ozdemir_guvenc_2020, title={Detection and Classification of UAVs Using RF Fingerprints in the Presence of Wi-Fi and Bluetooth Interference}, volume={1}, ISSN={["2644-125X"]}, url={https://doi.org/10.1109/OJCOMS.2019.2955889}, DOI={10.1109/OJCOMS.2019.2955889}, abstractNote={This paper investigates the problem of detection and classification of unmanned aerial vehicles (UAVs) in the presence of wireless interference signals using a passive radio frequency (RF) surveillance system. The system uses a multistage detector to distinguish signals transmitted by a UAV controller from the background noise and interference signals. First, RF signals from any source are detected using a Markov models-based naïve Bayes decision mechanism. When the receiver operates at a signal-to-noise ratio (SNR) of 10 dB, and the threshold, which defines the states of the models, is set at a level 3.5 times the standard deviation of the preprocessed noise data, a detection accuracy of 99.8% with a false alarm rate of 2.8% is achieved. Second, signals from Wi-Fi and Bluetooth emitters, if present, are detected based on the bandwidth and modulation features of the detected RF signal. Once the input signal is identified as a UAV controller signal, it is classified using machine learning (ML) techniques. Fifteen statistical features extracted from the energy transients of the UAV controller signals are fed to neighborhood component analysis (NCA), and the three most significant features are selected. The performance of the NCA and five different ML classifiers are studied for 15 different types of UAV controllers. A classification accuracy of 98.13% is achieved by k-nearest neighbor classifier at 25 dB SNR. Classification performance is also investigated at different SNR levels and for a set of 17 UAV controllers which includes two pairs from the same UAV controller models.}, journal={IEEE OPEN JOURNAL OF THE COMMUNICATIONS SOCIETY}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Ezuma, Martins and Erden, Fatih and Anjinappa, Chethan Kumar and Ozdemir, Ozgur and Guvenc, Ismail}, year={2020}, pages={60–76} }
 @article{khawaja_ozdemir_erden_ozturk_guvenc_2020, title={Multiple ray received power modelling for mmWave indoor and outdoor scenarios}, volume={14}, ISSN={["1751-8733"]}, url={https://doi.org/10.1049/iet-map.2020.0046}, DOI={10.1049/iet-map.2020.0046}, abstractNote={IET Microwaves, Antennas & PropagationVolume 14, Issue 14 p. 1825-1836 Research ArticleFree Access Multiple ray received power modelling for mmWave indoor and outdoor scenarios Wahab Khawaja, Corresponding Author Wahab Khawaja wahab.ali@must.edu.pk Department of Electrical and Computer Engineering, North Carolina State University, 890 Oval Dr, Raleigh, NC, 27606 USA Mirpur University of Science and Technology, Mirpur, AJK, PakistanSearch for more papers by this authorOzgur Ozdemir, Ozgur Ozdemir Department of Electrical and Computer Engineering, North Carolina State University, 890 Oval Dr, Raleigh, NC, 27606 USASearch for more papers by this authorFatih Erden, Fatih Erden orcid.org/0000-0002-1708-3063 Department of Electrical and Computer Engineering, North Carolina State University, 890 Oval Dr, Raleigh, NC, 27606 USASearch for more papers by this authorEnder Ozturk, Ender Ozturk orcid.org/0000-0002-6390-8089 Department of Electrical and Computer Engineering, North Carolina State University, 890 Oval Dr, Raleigh, NC, 27606 USASearch for more papers by this authorIsmail Guvenc, Ismail Guvenc Department of Electrical and Computer Engineering, North Carolina State University, 890 Oval Dr, Raleigh, NC, 27606 USASearch for more papers by this author Wahab Khawaja, Corresponding Author Wahab Khawaja wahab.ali@must.edu.pk Department of Electrical and Computer Engineering, North Carolina State University, 890 Oval Dr, Raleigh, NC, 27606 USA Mirpur University of Science and Technology, Mirpur, AJK, PakistanSearch for more papers by this authorOzgur Ozdemir, Ozgur Ozdemir Department of Electrical and Computer Engineering, North Carolina State University, 890 Oval Dr, Raleigh, NC, 27606 USASearch for more papers by this authorFatih Erden, Fatih Erden orcid.org/0000-0002-1708-3063 Department of Electrical and Computer Engineering, North Carolina State University, 890 Oval Dr, Raleigh, NC, 27606 USASearch for more papers by this authorEnder Ozturk, Ender Ozturk orcid.org/0000-0002-6390-8089 Department of Electrical and Computer Engineering, North Carolina State University, 890 Oval Dr, Raleigh, NC, 27606 USASearch for more papers by this authorIsmail Guvenc, Ismail Guvenc Department of Electrical and Computer Engineering, North Carolina State University, 890 Oval Dr, Raleigh, NC, 27606 USASearch for more papers by this author First published: 22 October 2020 https://doi.org/10.1049/iet-map.2020.0046Citations: 2AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract Millimetre-wave (mmWave) frequency bands are expected to be used for future fifth generation networks due to the availability of a large unused spectrum. However, the attenuation at mmWave frequencies is high. To resolve this issue, the utilisation of high gain antennas and beamforming mechanisms are widely investigated in the literature. In this work, the authors considered mmWave end-to-end propagation modelled by individual ray sources and explored the effects of the number of rays in the model and radiation patterns of the deployed antennas on the received power. It is shown that taking the dominant two rays is sufficient to model the channel for outdoor open areas as opposed to the indoor corridor which needs five dominant rays to have a good fit for the measurement and simulation results. It is observed that the radiation pattern of the antenna affects the slope of the path loss. Multi-path components increase the received power, thus, for indoor corridor scenarios, path loss according to the link distance is smaller for lower gain antennas due to increased reception of reflected components. For an outdoor open area, the slope of the path loss is found to be very close to that of the free space. 1 Introduction There has been a significant increase in the number of smart communication devices and high data rate applications in the last decade. This trend is expected to grow rapidly in the future [[1]]. However, the available spectrum at the sub-6 GHz band is limited. Higher frequency bands (e.g. millimetre-wave (mmWave) bands) are not heavily utilised, thus, offer larger bandwidths for wireless communication systems. Therefore, research efforts have been concentrated on exploring higher frequencies as an alternative to the sub-6 GHz band. The opening of the mmWave spectrum for mobile usage by FCC [[2]] has given a boost to the current research studies to best utilise these bands. However, mmWave communication suffers from its inherent high free space attenuation as well as high penetration losses. In this work, we used measurements, analytical ray modelling and ray-tracing simulations to model line-of-sight (LoS) characteristics of a mmWave communication channel in a corridor type indoor and open space outdoor environments at 28 GHz frequency band. We analytically calculated received signal properties using the dominant five-ray and two-ray received power models based on first-order reflections for the indoor corridor and outdoor open area, respectively. To compare with our analytical results, measurements were conducted at North Carolina State University using a PXI-based channel sounder platform from National Instruments, and two sets of directional horn antennas with gains 17 and 23 dBi at 28 GHz. The test setup used indoor and outdoor are shown in Figs. 1 and 2, respectively. Fig. 1Open in figure viewerPowerPoint Indoor corridor propagation setup at the basement of Engineering Building II, North Carolina State University Fig. 2Open in figure viewerPowerPoint Outdoor measurement setup at the top floor of a multi-storey car park, Centennial Campus, North Carolina State University The rest of the paper is organised as follows. Readers will find a comprehensive literature review as well as a summary of our contributions in Section 2. Section 3 includes details on received power modelling for indoor and outdoor environments. Section 4 covers experimental and ray-tracing simulations setup. In Section 5, the number of rays and percentage power sum of dominant five rays with a total power of rays is provided. In Section 6, results of measurements, simulations, and calculations for received power are given. In Section 7, a detailed discussion is presented for five-ray and two-ray models. Section 8 provides Ricean K-factor analysis and the paper ends with concluding remarks in Section 9. 2 Literature review and contributions Various approaches have been proposed in the literature to overcome the high attenuation problem at mmWave frequencies [[3], [4]]. A common method is to increase the gain or directivity of the antennas [[5], [6]]. The high directivity is obtained either by beamforming or deploying directional antennas (e.g. horn antennas). In addition to antenna type, material characteristics of the objects in the environment also play an important role in figuring the propagation statistics [[7], [8]]. One way of modelling propagation statistics is by using ray tracing. In the literature, different types of indoor geometries either in LoS or non-LoS (NLoS) scenarios for a wide variety of frequency bands are investigated using ray tracing software [[9]-[12]]. In this work, we modelled the end-to-end propagation as individual ray sources. For the indoor environment, five rays are used in calculations. One is the LoS and four are the reflected rays from two walls, ceiling, and ground. Each ray source contributes to the resulting received power. Contributions of the reflected rays are found to increase with the link distance. This is because when transmitter and receiver antennas are close, reflected fields are rejected by the receiver antenna because of its directional pattern. Together with high Fresnel reflection coefficient values along with the link, we observe an increase in the received power compared to free space, i.e. the slope of the path loss is smaller than that of the free space for indoor. For outdoor open area, two-ray model is found to be sufficient to model the received power and because of the absence of three first-order reflections, no obvious difference between path loss slopes have been detected. The analytical modelling results based on ray sources are compared with measurement and ray-tracing simulation results. We also made a comparative analysis of the measurements with five-ray and two-ray analytical models and ray-tracing simulations with five rays are provided for the indoor environment. The comparative analysis is carried out using z-test of the path loss model parameters. The z-test values indicate that the two-ray model does not provide a close match to the measurement path loss for the indoor corridor. On the other hand, the five-ray model provides a close fit to the measurement data. The ratio of power sum of dominant five rays to power sum of total rays obtained from measurements are also provided in this work. The percentage is greater than 90% for all the scenarios, which indicates that five rays are sufficient for modelling. The Ricean K-factor is also provided to study the contribution of LoS ray and diffuse rays over the link for two different gain antennas. Table 1 shows the related work in the literature, where ray tracing is used. Comparison of the available literature with our work highlights the following distinctions of our work: Propagation modelling based on dominant rays at 28 GHz is considered in our work. Table 1. Related work in the literature on mmWave channel modelling using ray tracing Literature Number of rays Frequency Maximum distance Reported channel statistics [[13]] 2 sub-6 GHz and mmWave 10 km Received power, two-ray model, break point distance based on first Fresnel zone [[14]] 1, 2, 5, 20 100 MHz, 1800 MHz, 2400 MHz 10 km Path loss, two-ray model, effect of first Fresnel zone on path loss exponent [[15]] 2 1.5 GHz 1 km Two-ray model, path loss exponent for vertically and horizontally polarised signals [[16]] 3 3.6 GHz, 10.6 GHz 100 m Path loss, three ray model for UWB propagation [[17]] 3 1900 MHz 400 m Three ray propagation model for PCS and -cellular services [[18]] 62 0.06 THz-1 THz 6 m Distance and frequency selective characteristics, coherence bandwidth, channel capacity, and temporal broadening analysis [[19]] 9 60 GHz 60 m LOS, 25 m NLOS Received power, indoor corridor power distribution comparison with Rayleigh and Rician [[20]] 2, 4, 5 2.4 GHz 50 m Received power analysis in open and closed corridors [[21]] 2, multiple rays 94 GHz 6 m Path loss, multipath analysis [[22]] 2, 4 94 'GHz 1.5 m Received power, multipath analysis for radars [[23]] 2, 4, 6, 10 2.4 GHz 10 m Path loss This study 2, 5 28 GHz 40 m indoor, 100 m outdoor Received power, path loss, the effect of antenna gain, adequacy analysis on number of rays using z-test and Ricean K-factor Five dominant rays were found to be adequate for the indoor corridor propagation modelling whereas, two dominant rays were found to adequately model the open area outdoors. The antenna gain of each individual ray is modelled based on its geometric position from the radiation pattern of the antenna provided in the datasheet. The resolvable distance of the rays compared to the LoS as a function of the link distance is also provided. Smaller than this resolvable distance, the rays will be superimposed coherently with the LoS component. A polarisation-dependent reflection coefficient for different materials is used at 28 GHz. A z-test is also performed for comparison of parameters of two-ray and five-ray path loss models obtained analytically, through ray-tracing simulations and measurements. A commonly occurring scenario for future fifth generation (5G) deployments is closely positioned transmitter and receivers at indoor corridors. This commonly occurring scenario in a typical indoor corridor environment is studied. There are other works in the literature in which five, even more, first-order reflections are taken into account [[18]-[27]]. We only considered rays experiencing the first-order reflection. This is because most of the received power comes from the LoS signal and the first-order reflections. For the purpose of illustration, consider that we have a second-order reflected ray with the same reflection coefficient as the first-order reflected ray, . Then the power contribution of the second-order reflected ray will be times the first order reflected ray. Generally, for common non-metallic surfaces such as walls and ground, hence, the power coming from second-order reflected ray will be smaller than the first-order reflected ray. In addition, rays that we consider as second-order reflections have to undergo longer paths than the first-order reflections as the geometry of a corridor obliges. Let the distance travelled by a first-order reflected ray from the transmitter to the receiver be , and additional distance travelled due to the second reflection. Then, the power of the second-order reflected ray will be smaller than the first-order reflected ray by a factor of due to free space path loss. Overall, the received power due to the second-order reflections will be significantly smaller compared to the first-order reflections and the contribution of the second-order reflections to the total received power will be small. Consequently, third and higher-order reflections will also have small contributions, thus negligible. The difference between first-order and second-order reflections in a similar setup is shown in a previous work, [[3]], as well. Moreover, considering higher-order reflections increases the complexity of the model unnecessarily compared to their contribution to the received power. Therefore, our model based on LoS and first-order reflections provides a robust and simple way to calculate received power in corridors and similar shaped indoor environments. Similarly, for outdoor open area two-ray model is sufficient to model the received power. 3 Received power modelling based on dominant rays for indoor corridor and outdoor open area In this section, we will first discuss antenna radiation pattern effects on propagation. Later, a received power calculation model based on dominant LoS signal and reflected rays in the indoor corridor (five-ray model) is presented. Two-ray model as a special case of the five-ray model is used for outdoor open area. 3.1 Antenna radiation pattern and propagation effects The antenna radiation pattern plays an important role in modelling the propagation characteristics of directional mmWave links. In the model, we used two directional horn antenna sets which have different gains and respective half-power beamwidths (HPBWs) in the azimuth and elevation planes. We represent the 3D antenna gain as a surface area extended on a sphere at a distance d with a given solid angle . The surface area A subtended by the antenna gain at a distance d from the source is , where the solid angle is given as: (1) where is the radiated power from the antenna in spherical coordinates as a function of distance d, elevation, and azimuth angles of and , respectively. is the maximum radiated power. The propagation from the transmitting antenna is modelled as a spherical wavefront. The majority of the radiated power is concentrated over the area covered by the solid angle represented by and , where these two angles represent the antenna HPBWs in the elevation and azimuth planes, respectively. Moreover, if and are small, we can approximate the area extended by and in space as at a fixed distance in the far-field region. The rays lying in this region will have significantly higher gain compared to the rays lying outside this area. 3.2 Ray resolution along the link distance The propagation from the antenna can be considered either as a wavefront propagation or decomposed as ray-based propagation [[28], [29]]. In the case of ray-based propagation, rays are considered to be originating from the transmitting antenna in all directions, where only the rays that interact with physical objects in the environment are taken into account. This is the basic principle of the ray tracing as well. In the case of a rectangular corridor, there are five dominant rays that interact with the surroundings. These rays are LOS, rays reflected from the left and right walls, floor, and ceiling. Theoretically, these rays are distinguishable at every point along the receiver route. The distinguishing characteristics of the individual rays depend on the (i) antenna characteristics at the transmitter and receiver; (ii) the geometry of propagation setup; and (iii) geometry of the environment. Our channel sounder setup can resolve any two rays at a spatial distance represented as , whereas the theoretical ray resolution can resolve rays at any distance. Consider the case of two-ray modelling for a given height of the transmitter and receiver represented as and , respectively (Fig. 3). When the link distance d between the transmitter and receiver is increased such that the difference between the paths travelled by any two rays is smaller than , those rays cannot be resolved, thus can be measured as a superposition. The relevant inequality is as follows: (2) Fig. 3Open in figure viewerPowerPoint Propagation of LoS and GRC from transmitter antenna towards receiver antenna when their heights ( and ) are the same Similarly, for the indoor corridor, the rays reflected from the ground, ceiling, and walls may not be resolvable depending on the link distance d. Fig. 4 shows the difference in path distances of the rays reflected from the ground, ceiling, and walls with respect to the LoS ray. In Fig. 4, the reflected rays are considered to be independent of each other. According to Fig. 4, the ray from the ceiling is the first to get unresolved at 3.1 m compared to ground reflected ray, which gets unresolved at 7 m. The rays from the two walls are not resolvable after 5 m. This indicates that the path of the reflected ray from the ceiling is the smallest compared to the paths of the remaining three rays. Fig. 4Open in figure viewerPowerPoint Difference of ray lengths with the LoS, plotted as a function of link distance 3.3 Received power modelling for indoor corridor The received signal is given by , where represents the transmitted signal, is the impulse response of the channel and is the convolution operation. In case that received and transmitted signals are known, channel impulse response (CIR) could be obtained by applying deconvolution. In this work, we considered the CIR in the indoor corridor (similar to rectangular waveguide). The reason for selecting a corridor is because the future 5G base stations (BSs)/access points (APs) are expected to be deployed in corridors of common building structures (e.g. buildings containing offices or classrooms). The major occupancy in these environments is in the rooms adjacent to the common corridor. Therefore, the BSs/APs are preferred in the corridor to provide optimum coverage to the adjoining rooms and in the corridor itself. Moreover, the indoor corridor can be considered to be a large rectangular room, or a square room (if considered in small portions). Therefore, propagation in this area can help to understand the propagation in other similar environments. The indoor corridor propagation layout is shown in Fig. 5. Based on the 3D geometry shown in Fig. 5, five dominant rays are considered. The characteristics of these rays depend on the antenna radiation pattern at the transmitter and receiver in both azimuth and elevation planes, height of the transmitter and receiver, and distance of the transmitter from the walls, ceiling, and floor for a given receiver position. The height of the transmitter and receiver are kept the same throughout the experiments. The five dominant rays are given as follows: one is the LoS and the other four are the reflected rays from the ground, ceiling, and two walls. Fig. 5Open in figure viewerPowerPoint Layout of the indoor corridor propagation environment As the distance of the receiver increases from the transmitter moving in a straight line received power coming from reflected rays increase as well. Due to the geometry of the test setup, as the link distance increases, reflected rays gets closer to the boresight of the received antenna, thus, captured with a higher gain. As a result of this, the difference between power value calculated taking only free space path loss into account and the five-ray received power increases in favour of the five-ray model. The contribution of the reflected rays to the overall received power is also dependent on the Fresnel reflection coefficients. Reflected rays of more than first-order have a significantly smaller contribution to the received power compared to first-order reflections. Therefore, in our model, we can safely ignore their contributions. Let represent the received LoS component given as: (3) where is the gain of the antenna for the transmitter at elevation and azimuth angles of and , respectively. Similarly, is the gain of the antenna for the receiver at elevation and azimuth angles of and , respectively, represents the delay of the LoS component given by , where c is the speed of the light and d is the distance of the LoS component, represents the phase of the LoS component, represents the dot product between the polarisation unit vectors of the electric field at the transmitter and receiver, respectively. The gain of the antenna for the LoS ray in the azimuth and elevation planes at the transmitter and receiver are given as follows [[30]]: (4) where and represent the direction of departure (DoD) in the elevation and azimuth planes, respectively. Similarly, the direction of arrival (DoA) in the elevation and azimuth planes are given as and . can be expressed as follows: (5) where is the antenna gain and is the relative phase of the component of a ray. If both the transmitter and receiver are aligned to their boresight, then the total gain given in (4) is maximised. Similar to the LoS component, the four dominant received rays reflected from the environment, with the ray index , is expressed as: (6) The reflection coefficient also called Fresnel reflection coefficient for the relative permittivity of the ground material is given as: (7) where the value of Y depends on the polarisation and are given for vertical and horizontal polarisation as follows: (8) If the link distance , then and the gain of the reflected ray approaches to the LoS component gain and the Fresnel reflection coefficient, . Let E represent the average over time, and represent the total received power, then , the coherent addition of the LoS and the reflected rays for , is given as: (9) Equation (9) can be rewritten for d values such that the reflected rays can be resolvable (see Section 3.2, Fig. 4) from each other: (10) From (3), (6), if , and where is the transmitted power. Moreover, for the LoS component, the XPD (cross polarisation discrimination) factor is negligible for vertical–vertical (VV) and horizontal–horizontal (HH) antenna orientations. Similarly, for the reflected rays, the diffuse scattering is small due to smooth reflecting surfaces leading to small XPDs. Therefore, the dot product of the polarisation vectors can be taken as 1 for the LoS and reflected rays. Therefore, the total received power from (9) can be written as follows: (11) where for . Additionally, if the heights of the antennas are not the same and/or not aligned to the boresight, we have additional attenuation due to smaller antenna gain. This attenuation will decrease with the increase in distance between the transmitter and the receiver. Considering the th individual reflected ray at a given link distance, we can write the received power as follows: (12) From (12), it can be observed that the received power of the th reflected ray approaches to the LoS ray at distance when (i) the antenna gains at the transmitter and receiver side are equal to the boresight antenna gains, and (ii) the reflection coefficient is 1. 3.4 Received power modelling for outdoor open area The outdoor open area is selected to study the mmWave propagation with two dominant rays, i.e. LOS and ground reflected component (GRC) with negligible contribution from the surroundings. Therefore, a simple propagation model that only considers two multipath components can accurately characterise the signal propagation. The outdoor open area scenario can be often encountered in parking lots, recreational parks, and highways, among others. The two-ray model can also be considered as a special case of the five-ray model. The two-ray model is used for received power modelling in outdoor open area assuming that antenna heights are significantly high. The contribution of any other rays from far off scatterers is small for the open area and is ignored. In the two-ray modelling, the received power is dependent on the LoS and GRC. Therefore, the total received power is given as follows: (13) where is the phase difference between the LoS and the GRC signals. 3.5 Polarisation effects on the received power The polarisation of electric fields should be taken into account. There are two co-polarised configurations based on antenna orientation used in the measurements, namely VV and HH. The difference in VV and HH antenna orientations is subject to the antenna radiation pattern in the azimuth and elevation planes. However, even though the whole patterns are different in two orthogonal planes, as the HPBWs are the same for both horn antenna sets, no significant difference in the antenna radiation patterns has been observed due to antenna orientation. Cross polarisation of vertical–horizontal (VH) is also introduced to study the XPD factor in the indoor corridor. Considering the channel stationary, we can obtain the XPD factor between the transmitter and receiver as follows: (14) where , and are the received powers for VV, VH and HH antenna orientations, respectively, and E denotes the expected value. A major use of XPD factor is that it helps to study the interaction of the antennas of different beamwidths with the surroundings when cross-polarisation is not negligible. 3.6 Path loss modelling The path loss obtained from the received power measured at different distances from the transmitter are given as follows: (15) An alpha–beta model for the path loss modelling [[31]] is given as: (16) where is the y-intercept in dB, is the slope and X is a random variable and , where expressed in dB is the variance of X. A least square regression is used to fit a regression line (best fit) to the data. 4 Experimental and ray-tracing simulations setup In this section, an indoor and outdoor experimental setup, as well as the ray-tracing simulation setup, are discussed. 4.1 Indoor and outdoor measurement setup Indoor corridor measurements were carried out at the basement of the Engineering Building II, North Carolina State University, shown in Fig. 1. The walls in the corridor are three-layered drywall, the ceiling is Armstrong type ceiling and the ground is a concrete grinded surface. The measurements were carried out using NI mmWave transceiver system operating at 28 GHz. The description of the NI mmWave transceiver system is provided in [[32]]. Two horn antenna sets with gains 17 and 23 dBi were used in the measurements. The HPBWs of 17 dBi antennas are and in the E- and H-planes, respectively. The HPBWs for the 23 dBi antennas in the E- and H-planes are and , respectively. The height of the transmitter and receiver from the ground was fixed to 1.44 m, whereas, the distance of the transmitter and receiver from the ceiling was 0.9 m. The distance from either of the walls to the antennas was 1.24 m. The transmitter was kept at a fixed position, whereas the receiver was moved in a straight line away from the transmitter at constant intervals of 0.3 m starting from 1.9 to 39.7 m. Laser alignment is used between the transmitter and the receiver at every step. The outdoor measurements were carried out at the top floor of a multi-storey car park at North Carolina State University shown in Fig. 2. Similar to the indoor corridor measurements, the transmitter was kept at a fixed place, and the receiver was moved in steps of 5 m beginning from 4.6 to 100 m. The height of the transmitter and receiver was 1.09 m. For both indoor and outdoor measurements, the transmit power has been set to 0 dBm. 4.2 Ray tracing and analytical simulation setup Ray-tracing simulations were carried out using Wireless InSite® software [[33]]. The environment model is shown in Fig. 6. The indoor corridor and the outdoor open area were modelled similar to the real environment with as many details as we could. For the indoor setup, four different material types are used for walls, floor, ceiling, and doors in the ray-tracing simulation environment. The material used for the floor is concrete, whereas, for the ceiling and walls, Armstrong ceiling and drywalls are used, respectively. The relative permittivity of the concrete floor at 28 GHz is 5.31, while it is 3 f}, number={14}, journal={IET MICROWAVES ANTENNAS & PROPAGATION}, publisher={Institution of Engineering and Technology (IET)}, author={Khawaja, Wahab and Ozdemir, Ozgur and Erden, Fatih and Ozturk, Ender and Guvenc, Ismail}, year={2020}, month={Nov}, pages={1825–1836} }
 @article{khawaja_ozdemir_erden_guvenc_matolak_2020, title={Ultra-Wideband Air-to-Ground Propagation Channel Characterization in an Open Area}, volume={56}, ISSN={["1557-9603"]}, url={https://doi.org/10.1109/TAES.2020.3003104}, DOI={10.1109/TAES.2020.3003104}, abstractNote={This article studies the air-to-ground ultra-wideband channel through propagation measurements between 3.1 to 4.8 GHz using unmanned-aerial-vehicles (UAVs). Different line-of-sight (LOS) and obstructed-LOS scenarios and two antenna orientations were used in the experiments. Multipath channel statistics for different propagation scenarios were obtained, and the Saleh–Valenzuela model was found to provide a good fit for the statistical channel model. An analytical path loss model based on antenna gains in the elevation plane is provided for unobstructed UAV hovering and moving (in a circular path) propagation scenarios.}, number={6}, journal={IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Khawaja, Wahab and Ozdemir, Ozgur and Erden, Fatih and Guvenc, Ismail and Matolak, David W.}, year={2020}, month={Dec}, pages={4533–4555} }
 @article{eroglu_erden_guvenc_2019, title={Adaptive Kalman Tracking for Indoor Visible Light Positioning}, volume={2019-November}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85082400593&partnerID=MN8TOARS}, DOI={10.1109/milcom47813.2019.9021016}, abstractNote={Visible light communication (VLC) utilizes light-emitting diodes (LEDs) to transmit wireless data. A VLC network can also be used to localize mobile users in indoor environments, where the global positioning system (GPS) signals are weak. However, the line-of-sight (LOS) links of mobile VLC devices can be blocked easily, which decreases the accuracy of localization. In this paper, we study tracking a VLC user when the availability of VLC access point (AP) link changes over the user's route. We propose a localization method for a single available AP and use known estimation methods when a larger number of APs are available. Tracking mobile users with Kalman filter can increase the accuracy of the positioning, but the generic Kalman filter does not consider instant changes in the measurement method. In order to include this information in the position estimation, we implement an adaptive Kalman filter by modifying the filter parameters based on the availability of APs to the user. Simulation results show that the implemented method decreases the root-mean-square error (RMSE) of the localization down to 30% −50% of the original estimation.}, journal={MILCOM 2019 - 2019 IEEE Military Communications Conference (MILCOM)}, publisher={IEEE}, author={Eroglu, Yusuf Said and Erden, Fatih and Guvenc, Ismail}, year={2019}, month={Sep} }
 @article{detection and classification of uavs using rf fingerprints in the presence of interference_2019, year={2019}, month={Sep} }
 @article{khawaja_ozdemir_erden_guvenc_ezuma_kakishima_2019, title={Effect of Passive Reflectors for Enhancing Coverage of 28 GHz mmWave Systems in an Outdoor Setting}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85068466022&partnerID=MN8TOARS}, DOI={10.1109/rws.2019.8714266}, abstractNote={The availability of large unused spectrum at millimeter wave (mmWave) frequency bands has steered the future 5G research towards these bands. However, mmWave signals are attenuated severely in the non-line-of-sight (NLOS) scenarios, thereby leaving the strong link quality by a large margin to line-of-sight (LOS) links. In this paper, a passive metallic reflector is used to enhance the coverage for mmWave signals in an outdoor, NLOS propagation scenarios. The received power from different azimuth and elevation angles are measured at 28 GHz in a parking lot setting. Our results show that using a 33 inch by 33 inch metallic reflector, the received power can be enhanced by 19 dB compared to no reflector case.}, journal={2019 IEEE Radio and Wireless Symposium (RWS)}, publisher={IEEE}, author={Khawaja, Wahab Ali Gulzar and Ozdemir, Ozgur and Erden, Fatih and Guvenc, Ismail and Ezuma, Martins and Kakishima, Yuichi}, year={2019}, month={Jan} }
 @article{ezuma_erden_anjinappa_ozdemir_guvenc_2019, title={Micro-UAV Detection and Classification from RF Fingerprints Using Machine Learning Techniques}, volume={2019-March}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85068315515&partnerID=MN8TOARS}, DOI={10.1109/aero.2019.8741970}, abstractNote={This paper focuses on the detection and classification of micro-unmanned aerial vehicles (UAVs)using radio frequency (RF)fingerprints of the signals transmitted from the controller to the micro-UAV. In the detection phase, raw signals are split into frames and transformed into the wavelet domain to remove the bias in the signals and reduce the size of data to be processed. A naive Bayes approach, which is based on Markov models generated separately for UAV and non-UAV classes, is used to check for the presence of a UAV in each frame. In the classification phase, unlike the traditional approaches that rely solely on time-domain signals and corresponding features, the proposed technique uses the energy transient signal. This approach is more robust to noise and can cope with different modulation techniques. First, the normalized energy trajectory is generated from the energy-time-frequency distribution of the raw control signal. Next, the start and end points of the energy transient are detected by searching for the most abrupt changes in the mean of the energy trajectory. Then, a set of statistical features is extracted from the energy transient. Significant features are selected by performing neighborhood component analysis (NCA)to keep the computational cost of the algorithm low. Finally, selected features are fed to several machine learning algorithms for classification. The algorithms are evaluated experimentally using a database containing 100 RF signals from each of 14 different UAV controllers. The signals are recorded wirelessly using a high-frequency oscilloscope. The data set is randomly partitioned into training and test sets for validation with the ratio 4:1. Ten Monte Carlo simulations are run and results are averaged to assess the performance of the methods. All the micro-UAVs are detected correctly and an average accuracy of 96.3% is achieved using the k-nearest neighbor (kNN)classification. Proposed methods are also tested for different signal-to-noise ratio (SNR)levels and results are reported.}, journal={2019 IEEE Aerospace Conference}, publisher={IEEE}, author={Ezuma, Martins and Erden, Fatih and Anjinappa, Chethan Kumar and Ozdemir, Ozgur and Guvenc, Ismail}, year={2019}, month={Mar} }
 @article{micro-uav detection and classification from rf fingerprints using machine learning techniques_2019, year={2019}, month={Jan} }
 @article{chowdhury_erden_guvenc_2019, title={RSS-Based Q-Learning for Indoor UAV Navigation}, volume={2019-November}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85082399878&partnerID=MN8TOARS}, DOI={10.1109/milcom47813.2019.9020894}, abstractNote={In this paper, we focus on the potential use of unmanned aerial vehicles (UAVs) for search and rescue (SAR) missions in GPS-denied indoor environments. We consider the problem of navigating a UAV to a wireless signal source, e.g., a smartphone or watch owned by a victim. We assume that the source periodically transmits RF signals to nearby wireless access points. Received signal strength (RSS) at the UAV, which is a function of the UAV and source positions, is fed to a Q-learning algorithm, and the UAV is navigated to the vicinity of the source. Unlike the traditional location-based Q-learning approach that uses the GPS coordinates of the agent, our method uses the RSS to define the states and rewards of the algorithm. It does not require any a priori information about the environment. These, in turn, make it possible to use the UAVs in indoor SAR operations. Two indoor scenarios with different dimensions are created using a ray tracing software. Then, the corresponding heat maps that show the RSS at each possible UAV location are extracted for more realistic analysis. Performance of the RSS-based Q-learning algorithm is compared with the baseline (location-based) Q-learning algorithm in terms of convergence speed, average number of steps per episode, and the total length of the final trajectory. Our results show that the RSS-based Q-learning provides competitive performance with the location-based Q-learning.}, journal={MILCOM 2019 - 2019 IEEE Military Communications Conference (MILCOM)}, publisher={IEEE}, author={Chowdhury, Md Moin Uddin and Erden, Fatih and Guvenc, Ismail}, year={2019}, month={May} }
 @article{khawaja_ozdemir_erden_guvenc_matolak_2019, title={UWB Air-to-Ground Propagation Channel Measurements and Modeling Using UAVs}, volume={2019-March}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85068349923&partnerID=MN8TOARS}, DOI={10.1109/aero.2019.8741964}, abstractNote={This paper presents an experimental study of the air-to-ground (AG) propagation channel through ultrawideband (UWB) measurements in an open area using unmanned-aerial-vehicles (UAVs). Measurements were performed using UWB radios operating in the frequency range of 3.1 GHz-4.8 GHz and UWB planar elliptical dipole antennas having an omni-directional pattern in the azimuth plane and typical donut shaped pattern in the elevation plane. Three scenarios were considered for the channel measurements: (i) two receivers (RXs) at different heights above the ground and placed close to each other in line-of-sight (LOS) with the transmitter (TX) on the UAV and the UAV is hovering; (ii) RXs in obstructed line-of-sight (OLOS) with the UAV TX due to foliage, and the UAV is hovering; and, (iii) UAV moving in a circular path. Different horizontal and vertical distances between the RXs and the TX were used in the measurements. In addition, two different antenna orientations were used on the UAV antennas (vertical and horizontal) to analyze the effects of antenna radiation patterns on the UWB AG propagation. From the empirical results, it was observed that the received power depends mainly on the antenna radiation pattern in the elevation plane when the antennas are oriented in the same direction, as expected for these omni-azimuth antennas. Moreover, the overall antenna gain at the TX and RX can be approximated using trigonometric functions of the elevation angle. The antenna orientation (polarization) mismatch increases path loss, and produces a larger number of weak multipath components (MPCs) than when aligned. Similarly, additional path loss and a larger number of MPCs were observed for the OLOS scenario. In the case of the UAV moving in a circular path, the antenna orientation mismatch has smaller effects on the path loss than when the UAV is hovering, because a larger number of cross polarized components are received during motion. A statistical channel model for UWB AG propagation is built from the empirical results.}, journal={2019 IEEE Aerospace Conference}, publisher={IEEE}, author={Khawaja, Wahab and Ozdemir, Ozgur and Erden, Fatih and Guvenc, Ismail and Matolak, David W.}, year={2019}, month={Mar} }
 @article{ultra-wideband air-to-ground propagation channel characterization in an open area_2019, year={2019}, month={Jun} }
 @article{vehicular lte connectivity analysis in urban and rural environments using usrp measurements_2019, year={2019}, month={Sep} }
 @article{erdogan_erden_kisacikoglu_2018, title={A fast and efficient coordinated vehicle-to-grid discharging control scheme for peak shaving in power distribution system}, volume={6}, url={http://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcAuth=ORCID&SrcApp=OrcidOrg&DestLinkType=FullRecord&DestApp=WOS_CPL&KeyUT=WOS:000431807000014&KeyUID=WOS:000431807000014}, DOI={10.1007/s40565-017-0375-z}, abstractNote={This study focuses on the potential role of plug-in electric vehicles (PEVs) as a distributed energy storage unit to provide peak demand minimization in power distribution systems. Vehicle-to-grid (V2G) power and currently available information transfer technology enables utility companies to use this stored energy. The V2G process is first formulated as an optimal control problem. Then, a two-stage V2G discharging control scheme is proposed. In the first stage, a desired level for peak shaving and duration for V2G service are determined off-line based on forecasted loading profile and PEV mobility model. In the second stage, the discharging rates of PEVs are dynamically adjusted in real time by considering the actual grid load and the characteristics of PEVs connected to the grid. The optimal and proposed V2G algorithms are tested using a real residential distribution transformer and PEV mobility data collected from field with different battery and charger ratings for heuristic user case scenarios. The peak shaving performance is assessed in terms of peak shaving index and peak load reduction. Proposed solution is shown to be competitive with the optimal solution while avoiding high computational loads. The impact of the V2G management strategy on the system loading at night is also analyzed by implementing an off-line charging scheduling algorithm.}, number={3}, journal={Journal of Modern Power Systems and Clean Energy}, author={Erdogan, N. and Erden, F. and Kisacikoglu, M.}, year={2018}, pages={555–566} }
 @article{erden_kisacikoglu_erdogan_2018, title={Adaptive V2G Peak Shaving and Smart Charging Control for Grid Integration of PEVs}, volume={46}, ISSN={["1532-5016"]}, url={https://doi.org/10.1080/15325008.2018.1489435}, DOI={10.1080/15325008.2018.1489435}, abstractNote={Abstract The stochastic nature of plug-in electric vehicle (PEV) driving behavior and distribution grid load profile make it challenging to control vehicle-grid integration in a mutually beneficial way. This article proposes a new adaptive control strategy that manages PEV charging/discharging for peak shaving and load leveling in a distribution grid. For accurate and high fidelity transportation mobility modeling, real vehicle driving test data are collected from the field. Considering the estimated total required PEV battery charging energy, the vehicle-to-grid capabilities of PEVs, and the forecasted non-PEV base load, a reference operating point for the grid is estimated. This reference operating point is updated once at the end of peak hours to guarantee a full final state-of-charge to each PEV. Proposed method provides cost-efficient operation for the utility grid, utmost user convenience free from range anxiety, and ease of implementation at the charging station nodes. It is tested on a real residential transformer, which serves approximately one thousand customers, under various PEV penetration levels and charging scenarios. Performance is assessed in terms of mean-square-error and peak shaving index. Results are compared with those of various reference operating point choices and shown to be superior.}, number={13}, journal={ELECTRIC POWER COMPONENTS AND SYSTEMS}, publisher={Informa UK Limited}, author={Erden, Fatih and Kisacikoglu, Mithat C. and Erdogan, Nuh}, year={2018}, month={Aug}, pages={1494–1508} }
 @inproceedings{ucer_kisacikoglu_erden_meintz_rames_2018, title={Development of a DC Fast Charging Station Model for use with EV Infrastructure Projection Tool}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85053825474&partnerID=MN8TOARS}, DOI={10.1109/ITEC.2018.8450158}, abstractNote={The deployment of public charging infrastructure networks has been a major factor in enabling electric vehicle (EV) technology transition, and must continue to support the adoption of this technology. DC fast charging (DCFC) increases customer convenience by lowering charging time, enables long-distance EV travel, and could allow the electrification of high-mileage fleets. Yet, high capital costs and uneven power demand have been major challenges to the widespread deployment of DCFC stations. There is a need to better understand DCFC stations’ loading, utilization, and customer service quality (i.e. queuing time, charging duration, and queue length). This study aims to analyze these aspects using one million vehicle-days of travel data within the Columbus, OH, region. Monte Carlo analysis is carried out in three types of areas - urban, suburban, and rural- to quantify the effect of uncertain parameters on DCFC station loading and service quality.}, booktitle={2018 IEEE Transportation and Electrification Conference and Expo, ITEC 2018}, author={Ucer, E.Y. and Kisacikoglu, M.C. and Erden, F. and Meintz, A. and Rames, C.}, year={2018}, pages={934–938} }
 @article{kisacikoglu_erden_erdogan_2018, title={Distributed Control of PEV Charging Based on Energy Demand Forecast}, volume={14}, url={http://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcAuth=ORCID&SrcApp=OrcidOrg&DestLinkType=FullRecord&DestApp=WOS_CPL&KeyUT=WOS:000422661900033&KeyUID=WOS:000422661900033}, DOI={10.1109/tii.2017.2705075}, abstractNote={This paper presents a new distributed smart charging strategy for grid integration of plug-in electric vehicles (PEVs). The main goal is to smooth the daily grid load profile while ensuring that each PEV has a desired state of charge level at the time of departure. Communication and computational overhead, and PEV user privacy are also considered during the development of the proposed strategy. It consists of two stages: 1) an offline process to estimate a reference operating power level based on the forecasted mobility energy demand and base loading profile, and 2) a real-time process to determine the charging power for each PEV so that the aggregated load tracks the reference loading level. Tests are carried out both on primary and secondary distribution networks for different heuristic charging scenarios and PEV penetration levels. Results are compared to that of the optimal solution and other state-of-the-art techniques in terms of variance and peak values, and shown to be competitive. Finally, a real vehicle test implementation is done using a commercial-of-the-shelf charging station and an electric vehicle.}, number={1}, journal={Ieee Transactions on Industrial Informatics}, author={Kisacikoglu, M. C. and Erden, F. and Erdogan, N.}, year={2018}, pages={332–341} }
 @article{hiranandani_mohadikar_khawaja_ozdemir_guvenc_matolak_2018, title={Effect of Passive Reflectors on the Coverage of IEEE 802.11ad mmWave Systems}, DOI={10.1109/vtcfall.2018.8690658}, abstractNote={Millimeter wave (mmWave) communications can provide high speed data rate for 5G wireless networks, taking advantage of the large bandwidth available at mmWave frequencies. It is also well known that mmWave links may suffer from outages due to non-line-of-sight (NLOS) propagation problems. In this paper, we study the characteristics of IEEE~802.11ad based mmWave systems at 60~GHz, and investigate the effect of passive reflectors for improving coverage. The experiments are performed using Tensorcom single chip TC60G-USB3-EVB and TC60G70100UP picocell platforms in different indoor and outdoor environments, both supporting beamforming based on IEEE~802.11ad specifications. Our results show that the passive metallic reflectors may signifciantly improve the mmWave signal coverage in regions where the line-of-sight (LOS) connectivity between the transmitter and the receiver is limited.}, journal={2018 IEEE 88th Vehicular Technology Conference (VTC-Fall)}, publisher={IEEE}, author={Hiranandani, Shivesh and Mohadikar, Sameer and Khawaja, Wahab and Ozdemir, Ozgur and Guvenc, Ismail and Matolak, David}, year={2018}, month={Aug} }
 @inproceedings{kisacikoglu_erden_erdogan_2017, title={A distributed smart PEV charging algorithm based on forecasted mobility energy demand}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85019250606&partnerID=MN8TOARS}, DOI={10.1109/GlobalSIP.2016.7905975}, abstractNote={This study proposes a new distributed control strategy for the grid integration of plug-in electric vehicles. The proposed strategy consists of two stages: (i) an offline process to determine an aggregated reference charge power level based on mobility estimation and base load profile, and (ii) a real-time operation based on the distributed control approach. The control algorithm manages PEV charge load profiles in order to flatten the residential distribution transformer loading while ensuring the desired state of the charge (SOC) level. The proposed algorithm is tested on real distribution transformer loading data, and compared with heuristic charging scenarios. The numerical results are presented to demonstrate the impact of the proposed algorithm.}, booktitle={2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings}, publisher={IEEE}, author={Kisacikoglu, Mithat C. and Erden, Fatih and Erdogan, Nuh}, year={2017}, pages={911–915} }
 @inproceedings{erden_cetin_2017, title={Breathing detection based on the topological features of IR sensor and accelerometer signals}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85016331834&partnerID=MN8TOARS}, DOI={10.1109/ACSSC.2016.7869685}, abstractNote={This paper describes a non-contact breathing detection system using a pyro-electric infrared (PIR) sensor and an accelerometer. The multi-sensor system can be used to detect the respiratory disorders. A PIR sensor is placed onto a stand near a bed and an accelerometer is placed on the mattress. We recently developed a PIR sensor which is capable of producing 1-D time-varying signals corresponding to the motions in its field of view. The PIR sensor signal due to the thoracic movements turns out to be an almost periodic signal. Similarly, the accelerometer produces an almost periodic signal in response to vibrations in bed. Sensor signals are processed using a topological approach. Point clouds are constructed from the delay-coordinate embedding of the time series sensor data first. Then, periodic structures in the point clouds are detected using persistent homology. The sensors, with the proposed method, complement each other to produce more accurate decisions in different lying positions.}, booktitle={Conference Record - Asilomar Conference on Signals, Systems and Computers}, publisher={IEEE}, author={Erden, Fatih and Cetin, A. Enis}, year={2017}, pages={1763–1767} }
 @article{erden_cetin_2017, title={IR Sensors for Indoor Monitoring}, DOI={10.1201/9781315155340-15}, abstractNote={This chapter focuses on infrared (IR) sensors that complement and have similar architectures to radars (both have transmitters/receivers). It presents the feature extraction methods and machine learning algorithms tailored for each specific application. Radars are compared to IR sensors for similar mode of operation highlighting the pros and cons of each. IR sensors have the following two types: active IR sensor; and passive IR (PIR) sensor. The chapter considers PIR sensor. A PIR sensor only has a collector to sense the motion of the subjects in its field of view (FOV). Rather than sensing the motion, the PIR sensor detects the change in the amount of IR radiation impinging upon it. PIR sensors will be used to refer to dual-element type PIR sensors unless otherwise stated. PIR sensor-based indoor monitoring applications can be categorized into the following fields: home security and safety, home automation, and health status monitoring.}, journal={Radar for Indoor Monitoring}, publisher={CRC Press}, author={Erden, Fatih and Cetin, A. Enis}, year={2017}, month={Sep}, pages={365–380} }
 @article{erden_cetin_2017, title={IR Sensors for Indoor Monitoring}, DOI={10.1201/9781315155340-16}, journal={Radar for Indoor Monitoring}, publisher={CRC Press}, author={Erden, Fatih and Cetin, A}, year={2017}, month={Oct}, pages={365–380} }
 @inbook{erden_cetin_2017, title={IR sensors for indoor monitoring}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85053333645&partnerID=MN8TOARS}, DOI={10.1201/9781315155340}, booktitle={Radar for Indoor Monitoring: Detection, Classification, and Assessment}, author={Erden, F. and Cetin, A.E.}, year={2017}, pages={365–380} }
 @article{erden_cetin_2017, title={Period Estimation of an Almost Periodic Signal Using Persistent Homology with Application to Respiratory Rate Measurement}, volume={24}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-85021729897&partnerID=MN8TOARS}, DOI={10.1109/LSP.2017.2699924}, abstractNote={Time-frequency techniques have difficulties in yielding efficient online algorithms for almost periodic signals. We describe a new topological method to find the period of signals that have an almost periodic waveform. Proposed method is applied to signals received from a pyro-electric infrared sensor array for the online estimation of the respiratory rate (RR) of a person. Time-varying analog signals captured from the sensors exhibit an almost periodic behavior due to repetitive nature of breathing activity. Sensor signals are transformed into two-dimensional point clouds with a technique that allows preserving the period information. Features, which represent the harmonic structures in the sensor signals, are detected by applying persistent homology and the RR is estimated based on the persistence barcode of the first Betti number. Experiments have been carried out to show that our method makes reliable estimates of the RR.}, number={7}, journal={IEEE Signal Processing Letters}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={erden and Cetin, A. Enis}, year={2017}, pages={1–1} }
 @inbook{erden_cent_2016, title={Breathing Detection Based on the Topological Features of IR Sensor and Accelerometer Signals}, url={http://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcAuth=ORCID&SrcApp=OrcidOrg&DestLinkType=FullRecord&DestApp=WOS_CPL&KeyUT=WOS:000406057400310&KeyUID=WOS:000406057400310}, booktitle={2016 50th Asilomar Conference on Signals, Systems and Computers}, author={Erden, F. and Cent, A. E.}, editor={Matthews, M. B.Editor}, year={2016}, pages={1763–1767} }
 @inproceedings{erden_kisacikoglu_gurec_2016, title={Examination of EV-grid integration using real driving and transformer loading data}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84963852502&partnerID=MN8TOARS}, DOI={10.1109/ELECO.2015.7394445}, abstractNote={The growing environmental concerns and the increase in oil prices will lead to the proliferation of electric vehicles (EVs) in the near future. The increase in the number of EVs, while providing green and inexpensive solutions to transportation needs, may cause constraints on the operation of the utility grid that should be investigated. In this paper, the real user driving information is collected from individual data tracking devices of passenger vehicle owners instead of assuming randomly distributed trip characteristics. The collected trip data are first analyzed to generate a statistical model of the trip characteristics in terms of home arrival times and state of charge (SOC) levels. The resulting model is then used to simulate and analyze the impact of EV integration in a real grid with different EV penetration levels. For this, real distribution transformer data provided by Başkent Electric Distribution Co. is used. The proposed method produces more realistic results in comparison to the studies assuming random scenarios.}, booktitle={ELECO 2015 - 9th International Conference on Electrical and Electronics Engineering}, publisher={IEEE}, author={Erden, Fatih and Kisacikoglu, Mithat C. and Gurec, Ozan H.}, year={2016}, pages={364–368} }
 @inproceedings{erden_cetin_2016, title={Respiratory rate monitoring using infrared sensors}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84963891368&partnerID=MN8TOARS}, DOI={10.1109/ELECO.2015.7394631}, abstractNote={Respiratory rate is an essential parameter in many practical applications such as patient and elderly people monitoring. In this paper, a novel contact-free system is introduced to detect the human breathing activity. The system, which consists of two pyro-electric infrared (PIR) sensors, is capable of estimating the respiratory rate and detecting the sleep apnea. Sensors' signals corresponding to the thoracic movements of a human being are sampled using a microprocessor and analyzed on a general-purpose computer. Sampled signals are processed using empirical mode decomposition (EMD) and a new average magnitude difference function (AMDF) is used to detect the periodicity and the period of the processed signals. The resulting period, by using the fact that breathing is almost a periodic activity, is monitored as the respiratory rate. The new AMDF provides a way to fuse the data from the multiple sensors and generate a more reliable estimation of the respiratory rate.}, booktitle={ELECO 2015 - 9th International Conference on Electrical and Electronics Engineering}, publisher={IEEE}, author={Erden, Fatih and Cetin, A. Enis}, year={2016}, pages={1136–1140} }
 @article{erden_velipasalar_alkar_cetin_2016, title={Sensors in Assisted Living A survey of signal and image processing methods}, volume={33}, url={http://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcAuth=ORCID&SrcApp=OrcidOrg&DestLinkType=FullRecord&DestApp=WOS_CPL&KeyUT=WOS:000372398000007&KeyUID=WOS:000372398000007}, DOI={10.1109/msp.2015.2489978}, abstractNote={Our society will face a notable demographic shift in the near future. According to a United Nations report, the ratio of the elderly population (aged 60 years or older) to the overall population increased from 9.2% in 1990 to 11.7% in 2013 and is expected to reach 21.1% by 2050 [1]. According to the same report, 40% of older people live independently in their own homes. This ratio is about 75% in the developed countries. These facts will result in many societal challenges as well as changes in the health-care system, such as an increase in diseases and health-care costs, a shortage of caregivers, and a rise in the number of individuals unable to live independently [2]. Thus, it is imperative to develop ambient intelligence-based assisted living (AL) tools that help elderly people live independently in their homes. The recent developments in sensor technology and decreasing sensor costs have made the deployment of various sensors in various combinations viable, including static setups as well as wearable sensors. This article presents a survey that concentrates on the signal processing methods employed with different types of sensors. The types of sensors covered are pyro-electric infrared (PIR) and vibration sensors, accelerometers, cameras, depth sensors, and microphones.}, number={2}, journal={Ieee Signal Processing Magazine}, author={Erden, Fatih and Velipasalar, Senem and Alkar, Ali Ziya and Cetin, A. Enis}, year={2016}, pages={36–44} }
 @article{erden_alkar_cetin_2015, title={A robust system for counting people using an infrared sensor and a camera}, volume={72}, url={http://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcAuth=ORCID&SrcApp=OrcidOrg&DestLinkType=FullRecord&DestApp=WOS_CPL&KeyUT=WOS:000362146700017&KeyUID=WOS:000362146700017}, DOI={10.1016/j.infrared.2015.07.019}, abstractNote={Abstract In this paper, a multi-modal solution to the people counting problem in a given area is described. The multi-modal system consists of a differential pyro-electric infrared (PIR) sensor and a camera. Faces in the surveillance area are detected by the camera with the aim of counting people using cascaded AdaBoost classifiers. Due to the imprecise results produced by the camera-only system, an additional differential PIR sensor is integrated to the camera. Two types of human motion: (i) entry to and exit from the surveillance area and (ii) ordinary activities in that area are distinguished by the PIR sensor using a Markovian decision algorithm. The wavelet transform of the continuous-time real-valued signal received from the PIR sensor circuit is used for feature extraction from the sensor signal. Wavelet parameters are then fed to a set of Markov models representing the two motion classes. The affiliation of a test signal is decided as the class of the model yielding higher probability. People counting results produced by the camera are then corrected by utilizing the additional information obtained from the PIR sensor signal analysis. With the proof of concept built, it is shown that the multi-modal system can reduce false alarms of the camera-only system and determines the number of people watching a TV set in a more robust manner.}, journal={Infrared Physics & Technology}, author={Erden, Fatih and Alkar, Ali Ziya and Cetin, Ahmet Enis}, year={2015}, pages={127–134} }
 @article{erden_alkar_cetin_2015, title={Contact-free measurement of respiratory rate using infrared and vibration sensors}, volume={73}, url={https://doi.org/10.1016/j.infrared.2015.09.005}, DOI={10.1016/j.infrared.2015.09.005}, abstractNote={Respiratory rate is an essential parameter in many practical applications such as apnea detection, patient monitoring, and elderly people monitoring. In this paper, we describe a novel method and a contact-free multi-modal system which is capable of detecting human breathing activity. The multimodal system, which uses both differential pyro-electric infrared (PIR) and vibration sensors, can also estimate the respiratory rate. Vibration sensors pick up small vibrations due to the breathing activity. Similarly, PIR sensors pick up the thoracic movements. Sensor signals are sampled using a microprocessor board and analyzed on a laptop computer. Sensor signals are processed using wavelet analysis and empirical mode decomposition (EMD). Since breathing is almost periodic, a new multi-modal average magnitude difference function (AMDF) is used to detect the periodicity and the period in the processed signals. By fusing the data of two different types of sensors we achieve a more robust and reliable contact-free human breathing activity detection system compared to systems using only one specific type of sensors.}, journal={Infrared Physics & Technology}, publisher={Elsevier BV}, author={Erden, Fatih and Alkar, Ali Ziya and Cetin, Ahmet Enis}, year={2015}, month={Nov}, pages={88–94} }
 @article{erden_kisacikoglu_gurec_ieee_2015, title={Examination of EV-Grid Integration Using Real Driving and Transformer Loading Data}, url={http://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcAuth=ORCID&SrcApp=OrcidOrg&DestLinkType=FullRecord&DestApp=WOS_CPL&KeyUT=WOS:000380410800066&KeyUID=WOS:000380410800066}, journal={2015 9th International Conference on Electrical and Electronics Engineering (Eleco)}, author={Erden, F. and Kisacikoglu, M. C. and Gurec, O. H. and Ieee}, year={2015}, pages={364–368} }
 @article{erden_cetin_ieee_2015, title={Respiratory Rate Monitoring Using Infrared Sensors}, url={http://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcAuth=ORCID&SrcApp=OrcidOrg&DestLinkType=FullRecord&DestApp=WOS_CPL&KeyUT=WOS:000380410800196&KeyUID=WOS:000380410800196}, journal={2015 9th International Conference on Electrical and Electronics Engineering (Eleco)}, author={Erden, F. and Cetin, A. E. and Ieee}, year={2015}, pages={1136–1140} }
 @inbook{erden_bingol_cetin_ieee_2014, title={HAND GESTURE RECOGNITION USING TWO DIFFERENTIAL PIR SENSORS AND A CAMERA}, url={http://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcAuth=ORCID&SrcApp=OrcidOrg&DestLinkType=FullRecord&DestApp=WOS_CPL&KeyUT=WOS:000356351400067&KeyUID=WOS:000356351400067}, booktitle={2014 22nd Signal Processing and Communications Applications Conference}, author={Erden, F. and Bingol, A. S. and Cetin, A. E. and Ieee}, year={2014}, pages={349–352} }
 @article{erden_cetin_2014, title={Hand gesture based remote control system using infrared sensors and a camera}, volume={60}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84923657192&partnerID=MN8TOARS}, DOI={10.1109/tce.2014.7027342}, abstractNote={In this paper, a multimodal hand gesture detection and recognition system using differential Pyroelectric Infrared (PIR) sensors and a regular camera is described. Any movement within the viewing range of the differential PIR sensors are first detected by the sensors and then checked if it is due to a hand gesture or not by video analysis. If the movement is due to a hand, one-dimensional continuous-time signals extracted from the PIR sensors are used to classify/recognize the hand movements in real-time. Classification of different hand gestures by using the differential PIR sensors is carried out by a new winner-takeall (WTA) hash based recognition method. Jaccard distance is used to compare the WTA hash codes extracted from 1-D differential infrared sensor signals. It is experimentally shown that the multimodal system achieves higher recognition rates than the system based on only the on/off decisions of the analog circuitry of the PIR sensors.}, number={4}, journal={IEEE Transactions on Consumer Electronics}, publisher={Institute of Electrical and Electronics Engineers (IEEE)}, author={Erden, Fatih and Cetin, A. Enis}, year={2014}, month={Nov}, pages={675–680} }
 @article{erden_bingol_cetin_2014, title={Hand gesture recognition using two differential PIR sensors and a camera}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84903755066&partnerID=MN8TOARS}, DOI={10.1109/siu.2014.6830237}, abstractNote={In this paper, a hand gesture detection and classification system using two differential Pyro-electric Infrared (PIR) sensors and a camera is introduced. Motion presence is investigated in the area of interest using two differential PIR sensors. In the case of any motion detection, a decision whether it belongs to a hand or not is made by using camera and if it is a hand, which gesture belongs to which predefined class is determined by evaluating each system data together. In the stage of hand gesture detection and classification using camera, skin detection and convex hull-defect algorithms are used. Classification of different hand gestures by using differential PIR sensors is carried out by Winner-Take-All (WTA) hash method. The main contribution of this paper is to show that WTA hash codes can be utilized in classification of 1-D signals and gesture recognition accuracy can be improved by multi-sensor fusion.}, journal={2014 22nd Signal Processing and Communications Applications Conference (SIU)}, publisher={IEEE}, author={Erden, F. and Bingol, A. S. and Cetin, A. E.}, year={2014}, month={Apr}, pages={349–352} }
 @book{dimitropoulos_gunay_kose_erden_chaabene_tsalakanidou_grammalidis_cetin_2012, title={Flame detection for video-based early fire warning for the protection of cultural heritage}, volume={7616 LNCS}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84868030366&partnerID=MN8TOARS}, DOI={10.1007/978-3-642-34234-9_38}, abstractNote={Cultural heritage and archaeological sites are exposed to the risk of fire and early warning is the only way to avoid losses and damages. The use of terrestrial systems, typically based on video cameras, is currently the most promising solution for advanced automatic wildfire surveillance and monitoring. Video cameras are sensitive in visible spectra and can be used either for flame or smoke detection. This paper presents and compares three video-based flame detection techniques, which were developed within the FIRESENSE EU research project.}, journal={Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)}, author={Dimitropoulos, K. and Gunay, O. and Kose, K. and Erden, F. and Chaabene, F. and Tsalakanidou, F. and Grammalidis, N. and Cetin, E.}, year={2012}, pages={378–387} }
 @article{erden_toreyin_soyer_inac_gunay_kose_cetin_2012, title={Wavelet based flame detection using differential PIR sensors}, url={http://www.scopus.com/inward/record.url?eid=2-s2.0-84863434890&partnerID=MN8TOARS}, DOI={10.1109/siu.2012.6204529}, abstractNote={In this paper, a flame detection system using a differential Pyro-electric Infrared (PIR) sensor is proposed. A differential PIR sensor is only sensitive to sudden temperature variations within its viewing range and it produces a time-varying signal. The wavelet transform of the differential PIR sensor signal is used for feature extraction and feature vectors are fed to Markov models trained with uncontrolled fire flames and walking person. The model yielding the highest probability is chosen. Results suggest that the system can be used in spacious rooms for uncontrolled fire flame detection.}, journal={2012 20th Signal Processing and Communications Applications Conference (SIU)}, publisher={IEEE}, author={Erden, F. and Toreyin, B.U. and Soyer, E.B. and Inac, I. and Gunay, O. and Kose, K. and Cetin, A.E.}, year={2012}, month={Apr} }
 @article{erden_toreyin_soyer_inac_gunay_kose_cetin_2012, title={Wavelet based flickering flame detector using differential PIR sensors}, volume={53}, url={http://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcAuth=ORCID&SrcApp=OrcidOrg&DestLinkType=FullRecord&DestApp=WOS_CPL&KeyUT=WOS:000309029500002&KeyUID=WOS:000309029500002}, DOI={10.1016/j.firesaf.2012.06.006}, abstractNote={A Pyro-electric Infrared (PIR) sensor based flame detection system is proposed using a Markovian decision algorithm. A differential PIR sensor is only sensitive to sudden temperature variations within its viewing range and it produces a time-varying signal. The wavelet transform of the PIR sensor signal is used for feature extraction from sensor signal and wavelet parameters are fed to a set of Markov models corresponding to the flame flicker process of an uncontrolled fire, ordinary activity of human beings and other objects. The final decision is reached based on the model yielding the highest probability among others. Comparative results show that the system can be used for fire detection in large rooms.}, journal={Fire Safety Journal}, author={Erden, Fatih and Toreyin, B. Ugur and Soyer, E. Birey and Inac, Ihsan and Gunay, Osman and Kose, Kivanc and Cetin, A. Enis}, year={2012}, pages={13–18} }
 @article{erden_soyer_toreyin_cetin_ieee_2010, title={VOC GAS LEAK DETECTION USING PYRO-ELECTRIC INFRARED SENSORS}, url={http://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcAuth=ORCID&SrcApp=OrcidOrg&DestLinkType=FullRecord&DestApp=WOS_CPL&KeyUT=WOS:000287096001167&KeyUID=WOS:000287096001167}, DOI={10.1109/ICASSP.2010.5495500}, abstractNote={In this paper, we propose a novel method for detecting and monitoring Volatile Organic Compounds (VOC) gas leaks by using a Pyro-electric (or Passive) Infrared (PIR) sensor whose spectral range intersects with the absorption bands of VOC gases. A continuous time analog signal is obtained from the PIR sensor. This signal is discretized and analyzed in real time. Feature parameters are extracted in wavelet domain and classified using a Markov Model (MM) based classifier. Experimental results are presented.}, journal={2010 Ieee International Conference on Acoustics, Speech, and Signal Processing}, publisher={IEEE}, author={Erden, Fatih and Soyer, E. Birey and Toreyin, B. Ugur and Cetin, A. Enis and IEEE}, year={2010}, pages={1682–1685} }