martins Ezuma | Inha University (original) (raw)

Papers by martins Ezuma

arXiv (Cornell University), Nov 18, 2018

The availability of large unused spectrum at millimeter wave (mmWave) frequency bands has steered... more 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-lineof-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.

arXiv (Cornell University), May 12, 2019

arXiv (Cornell University), Sep 11, 2019

arXiv (Cornell University), Feb 23, 2021

arXiv (Cornell University), Feb 8, 2024

arXiv (Cornell University), Nov 18, 2022

The use of unmanned aerial vehicles (UAVs) for different applications has increased many folds in... more The use of unmanned aerial vehicles (UAVs) for different applications has increased many folds in recent years. The UAVs are expected to change the future air operations. However, there are instances where the UAVs can be used for malicious purposes. The detection, tracking, and classification of UAVs is challenging compared to manned aerial vehicles (MAVs) mainly due to small size, complex shapes, and ability to fly close to the terrain and in autonomous flight patterns in swarms. In this survey, we will discuss current and future aerial threats, and provide an overview of radar systems to counter such threats. We also study the performance parameters of radar systems for the detection, tracking, and classification of UAVs compared to MAVs. In addition to dedicated radar systems, we review the use of joint communication-radar (JCR) systems, as well as passive monitoring of changes in the common communication signals, e.g., FM, LTE, and any transmissions that may radiate from a UAV, for the detection, tracking, and classification of UAVs are provided. Finally, limitations of radar systems and comparison with other techniques that do not rely on radars for detection, tracking, and classification of aerial threats are provided.

arXiv (Cornell University), Feb 23, 2021

• To detect the presence of radio-controlled UAVs in an environment by exploiting the RF signal e... more • To detect the presence of radio-controlled UAVs in an environment by exploiting the RF signal emanating from the UAV-flight controller communication under wireless interference (i.e., WiFi and Bluetooth). • To explore the possibility of extracting RF fingerprints from the transient and steady state of the RF signals for detection and identification of UAVs. • To utilize wavelet transform analytics (i.e., continuous wavelet transform and wavelet scattering transform) for the feature extraction where both coefficients and image-based signature are generated for training machine learning algorithms and convolutional neural network. • To evaluate the performance of trained models under varying signal to noise ratio.

arXiv (Cornell University), Nov 13, 2019

The knowledge of the radar signature of aerial targets, such as drones, is critical in designing ... more The knowledge of the radar signature of aerial targets, such as drones, is critical in designing an effective radar detection system. It is a challenging task to measure the radar cross-section (RCS) of small drones. This paper describes a compact-range approach for measuring the RCS of small drones at 15 GHz and 25 GHz. The measurement results show that the average RCS of the three small drones varies with the radar frequency with higher reflections observed around certain directions. Moreover, the results show that for each drone, the RCS at 25 GHz is higher than the RCS at 15 GHz. Besides, information-theoretical based model selection for the RCS data is carried using the Akaike information criterion (AIC). We find that the generalized extreme value distribution is a good fit for modeling the RCS of small drones.

arXiv (Cornell University), Dec 17, 2021

This work investigates the problem of unmanned aerial vehicles (UAVs) identification using their ... more This work investigates the problem of unmanned aerial vehicles (UAVs) identification using their radar crosssection (RCS) signature. The RCS of six commercial UAVs are measured at 15 GHz and 25 GHz in an anechoic chamber, for both vertical-vertical and horizontal-horizontal polarization. The RCS signatures are used to train 15 different classification algorithms, each belonging to one of three different categories: statistical learning (SL), machine learning (ML), and deep learning (DL). The study shows that while the classification accuracy of all the algorithms increases with the signal-tonoise ratio (SNR), the ML algorithm achieved better accuracy than the SL and DL algorithms. For example, the classification tree ML achieves an accuracy of 98.66% at 3 dB SNR using the 15 GHz VV-polarized RCS test data from the UAVs. We investigate the classification accuracy using Monte Carlo analysis with the aid of boxplots, confusion matrices, and classification plots. On average, the accuracy of the classification tree ML model performed better than the other algorithms, followed by the Peter Swerling statistical models and the discriminant analysis ML model. In general, the classification accuracy of the ML and SL algorithms outperformed the DL algorithms (Squeezenet, Googlenet, Nasnet, and Resnet 101) considered in the study. Furthermore, the computational time of each algorithm is analyzed. The study concludes that while the SL algorithms achieved good classification accuracy, the computational time was relatively long when compared to the ML and DL algorithms. Also, the study shows that the classification tree achieved the fastest average classification time of about 0.46 ms. Index Terms-Deep learning (DL), machine learning (ML), radar cross-section (RCS), statistical learning (SL), target classification and recognition, unmanned aerial vehicles (UAVs).

This work investigates the problem of unmanned aerial vehicles (UAVs) identification using their ... more This work investigates the problem of unmanned aerial vehicles (UAVs) identification using their radar crosssection (RCS) signature. The RCS of six commercial UAVs are measured at 15 GHz and 25 GHz in an anechoic chamber, for both vertical-vertical and horizontal-horizontal polarization. The RCS signatures are used to train 15 different classification algorithms, each belonging to one of three different categories: statistical learning (SL), machine learning (ML), and deep learning (DL). The study shows that while the classification accuracy of all the algorithms increases with the signal-tonoise ratio (SNR), the ML algorithm achieved better accuracy than the SL and DL algorithms. For example, the classification tree ML achieves an accuracy of 98.66% at 3 dB SNR using the 15 GHz VV-polarized RCS test data from the UAVs. We investigate the classification accuracy using Monte Carlo analysis with the aid of boxplots, confusion matrices, and classification plots. On average, the accurac...

The use of supervised learning with various sensing techniques such as audio, visual imaging, the... more The use of supervised learning with various sensing techniques such as audio, visual imaging, thermal sensing, RADAR, and radio frequency (RF) have been widely applied in the detection of unmanned aerial vehicles (UAV) in an environment. However, little or no attention has been given to the application of unsupervised or semi-supervised algorithms for UAV detection. In this paper, we proposed a semi-supervised technique and architecture for detecting UAVs in an environment by exploiting the RF signals (i.e., fingerprints) between a UAV and its flight-controller communication under wireless inference such as Bluetooth and WiFi. By decomposing the RF signals using a two-level wavelet packet transform, we estimated the second moment statistic (i.e., variance) of the coefficients in each packet as a feature set. We developed a local outlier factor model as the UAV detection algorithm using the coefficient variances of the wavelet packets from WiFi and Bluetooth signals. When detecting t...

• To detect the presence of radio-controlled UAVs in an environment by exploiting the RF signal e... more • To detect the presence of radio-controlled UAVs in an environment by exploiting the RF signal emanating from the UAV-flight controller communication under wireless interference (i.e., WiFi and Bluetooth). • To explore the possibility of extracting RF fingerprints from the transient and steady state of the RF signals for detection and identification of UAVs. • To utilize wavelet transform analytics (i.e., continuous wavelet transform and wavelet scattering transform) for the feature extraction where both coefficients and image-based signature are generated for training machine learning algorithms and convolutional neural network. • To evaluate the performance of trained models under varying signal to noise ratio.

IEEE Transactions on Aerospace and Electronic Systems, 2022

2019 IEEE Radio and Wireless Symposium (RWS), 2019

This dataset contains RF signals from drone remote controllers (RCs) of different makes and model... more This dataset contains RF signals from drone remote controllers (RCs) of different makes and models. The RF signals transmitted by the drone RCs to communicate with the drones are intercepted and recorded by a passive RF surveillance system, which consists of a high-frequency oscilloscope, directional grid antenna, and low-noise power amplifier. The drones were idle during the data capture process. All the drone RCs transmit signals in the 2.4 GHz band. There are 17 drone RCs from eight different manufacturers and ~1000 RF signals per drone RC, each spanning a duration of 0.25 ms.

2021 IEEE Aerospace Conference (50100), 2021

We propose a security architecture that uses the principle of chaos for UAV secure communication.... more We propose a security architecture that uses the principle of chaos for UAV secure communication. A UAV, identified as an aerial base station (ABS), communicates with a ground base station (GBS) over a wireless radio frequency (RF) channel. The communication units of the ABS and GBS have dynamics according to the logistic map. The map is chaotic in the appropriate parameter space. Its states are non-periodic, broadband, and noise-like in the frequency domain. They

arXiv: Signal Processing, 2019

This paper investigates the problem of detection and classification of unmanned aerial vehicles (... more 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 naive 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 lea...

arXiv: Signal Processing, 2018

The future 5G networks are expected to use millimeter wave (mmWave) frequency bands, mainly due t... more The future 5G networks are expected to use millimeter wave (mmWave) frequency bands, mainly due to the availability of large unused spectrum. However, due to high path loss at mmWave frequencies, coverage of mmWave signals can get severely reduced, especially for non-line-of-sight (NLOS) scenarios. In this work, we study the use of passive metallic reflectors of different shapes/sizes to improve mmWave signal coverage for indoor NLOS scenarios. Software defined radio based mmWave transceiver platforms operating at 28 GHz are used for indoor measurements. Subsequently, ray tracing (RT) simulations are carried out in a similar environment using Remcom Wireless InSite software. The cumulative distribution functions of the received signal strength for the RT simulations in the area of interest are observed to be reasonably close with those obtained from the measurements. Our measurements and RT simulations both show that there is significant (on the order of 20 dB) power gain obtained w...

2021 IEEE 32nd Annual International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), 2021

arXiv (Cornell University), Nov 18, 2018

The availability of large unused spectrum at millimeter wave (mmWave) frequency bands has steered... more 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-lineof-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.

arXiv (Cornell University), May 12, 2019

arXiv (Cornell University), Sep 11, 2019

arXiv (Cornell University), Feb 23, 2021

arXiv (Cornell University), Feb 8, 2024

arXiv (Cornell University), Nov 18, 2022

The use of unmanned aerial vehicles (UAVs) for different applications has increased many folds in... more The use of unmanned aerial vehicles (UAVs) for different applications has increased many folds in recent years. The UAVs are expected to change the future air operations. However, there are instances where the UAVs can be used for malicious purposes. The detection, tracking, and classification of UAVs is challenging compared to manned aerial vehicles (MAVs) mainly due to small size, complex shapes, and ability to fly close to the terrain and in autonomous flight patterns in swarms. In this survey, we will discuss current and future aerial threats, and provide an overview of radar systems to counter such threats. We also study the performance parameters of radar systems for the detection, tracking, and classification of UAVs compared to MAVs. In addition to dedicated radar systems, we review the use of joint communication-radar (JCR) systems, as well as passive monitoring of changes in the common communication signals, e.g., FM, LTE, and any transmissions that may radiate from a UAV, for the detection, tracking, and classification of UAVs are provided. Finally, limitations of radar systems and comparison with other techniques that do not rely on radars for detection, tracking, and classification of aerial threats are provided.

arXiv (Cornell University), Feb 23, 2021

• To detect the presence of radio-controlled UAVs in an environment by exploiting the RF signal e... more • To detect the presence of radio-controlled UAVs in an environment by exploiting the RF signal emanating from the UAV-flight controller communication under wireless interference (i.e., WiFi and Bluetooth). • To explore the possibility of extracting RF fingerprints from the transient and steady state of the RF signals for detection and identification of UAVs. • To utilize wavelet transform analytics (i.e., continuous wavelet transform and wavelet scattering transform) for the feature extraction where both coefficients and image-based signature are generated for training machine learning algorithms and convolutional neural network. • To evaluate the performance of trained models under varying signal to noise ratio.

arXiv (Cornell University), Nov 13, 2019

The knowledge of the radar signature of aerial targets, such as drones, is critical in designing ... more The knowledge of the radar signature of aerial targets, such as drones, is critical in designing an effective radar detection system. It is a challenging task to measure the radar cross-section (RCS) of small drones. This paper describes a compact-range approach for measuring the RCS of small drones at 15 GHz and 25 GHz. The measurement results show that the average RCS of the three small drones varies with the radar frequency with higher reflections observed around certain directions. Moreover, the results show that for each drone, the RCS at 25 GHz is higher than the RCS at 15 GHz. Besides, information-theoretical based model selection for the RCS data is carried using the Akaike information criterion (AIC). We find that the generalized extreme value distribution is a good fit for modeling the RCS of small drones.

arXiv (Cornell University), Dec 17, 2021

This work investigates the problem of unmanned aerial vehicles (UAVs) identification using their ... more This work investigates the problem of unmanned aerial vehicles (UAVs) identification using their radar crosssection (RCS) signature. The RCS of six commercial UAVs are measured at 15 GHz and 25 GHz in an anechoic chamber, for both vertical-vertical and horizontal-horizontal polarization. The RCS signatures are used to train 15 different classification algorithms, each belonging to one of three different categories: statistical learning (SL), machine learning (ML), and deep learning (DL). The study shows that while the classification accuracy of all the algorithms increases with the signal-tonoise ratio (SNR), the ML algorithm achieved better accuracy than the SL and DL algorithms. For example, the classification tree ML achieves an accuracy of 98.66% at 3 dB SNR using the 15 GHz VV-polarized RCS test data from the UAVs. We investigate the classification accuracy using Monte Carlo analysis with the aid of boxplots, confusion matrices, and classification plots. On average, the accuracy of the classification tree ML model performed better than the other algorithms, followed by the Peter Swerling statistical models and the discriminant analysis ML model. In general, the classification accuracy of the ML and SL algorithms outperformed the DL algorithms (Squeezenet, Googlenet, Nasnet, and Resnet 101) considered in the study. Furthermore, the computational time of each algorithm is analyzed. The study concludes that while the SL algorithms achieved good classification accuracy, the computational time was relatively long when compared to the ML and DL algorithms. Also, the study shows that the classification tree achieved the fastest average classification time of about 0.46 ms. Index Terms-Deep learning (DL), machine learning (ML), radar cross-section (RCS), statistical learning (SL), target classification and recognition, unmanned aerial vehicles (UAVs).

This work investigates the problem of unmanned aerial vehicles (UAVs) identification using their ... more This work investigates the problem of unmanned aerial vehicles (UAVs) identification using their radar crosssection (RCS) signature. The RCS of six commercial UAVs are measured at 15 GHz and 25 GHz in an anechoic chamber, for both vertical-vertical and horizontal-horizontal polarization. The RCS signatures are used to train 15 different classification algorithms, each belonging to one of three different categories: statistical learning (SL), machine learning (ML), and deep learning (DL). The study shows that while the classification accuracy of all the algorithms increases with the signal-tonoise ratio (SNR), the ML algorithm achieved better accuracy than the SL and DL algorithms. For example, the classification tree ML achieves an accuracy of 98.66% at 3 dB SNR using the 15 GHz VV-polarized RCS test data from the UAVs. We investigate the classification accuracy using Monte Carlo analysis with the aid of boxplots, confusion matrices, and classification plots. On average, the accurac...

The use of supervised learning with various sensing techniques such as audio, visual imaging, the... more The use of supervised learning with various sensing techniques such as audio, visual imaging, thermal sensing, RADAR, and radio frequency (RF) have been widely applied in the detection of unmanned aerial vehicles (UAV) in an environment. However, little or no attention has been given to the application of unsupervised or semi-supervised algorithms for UAV detection. In this paper, we proposed a semi-supervised technique and architecture for detecting UAVs in an environment by exploiting the RF signals (i.e., fingerprints) between a UAV and its flight-controller communication under wireless inference such as Bluetooth and WiFi. By decomposing the RF signals using a two-level wavelet packet transform, we estimated the second moment statistic (i.e., variance) of the coefficients in each packet as a feature set. We developed a local outlier factor model as the UAV detection algorithm using the coefficient variances of the wavelet packets from WiFi and Bluetooth signals. When detecting t...

• To detect the presence of radio-controlled UAVs in an environment by exploiting the RF signal e... more • To detect the presence of radio-controlled UAVs in an environment by exploiting the RF signal emanating from the UAV-flight controller communication under wireless interference (i.e., WiFi and Bluetooth). • To explore the possibility of extracting RF fingerprints from the transient and steady state of the RF signals for detection and identification of UAVs. • To utilize wavelet transform analytics (i.e., continuous wavelet transform and wavelet scattering transform) for the feature extraction where both coefficients and image-based signature are generated for training machine learning algorithms and convolutional neural network. • To evaluate the performance of trained models under varying signal to noise ratio.

IEEE Transactions on Aerospace and Electronic Systems, 2022

2019 IEEE Radio and Wireless Symposium (RWS), 2019

This dataset contains RF signals from drone remote controllers (RCs) of different makes and model... more This dataset contains RF signals from drone remote controllers (RCs) of different makes and models. The RF signals transmitted by the drone RCs to communicate with the drones are intercepted and recorded by a passive RF surveillance system, which consists of a high-frequency oscilloscope, directional grid antenna, and low-noise power amplifier. The drones were idle during the data capture process. All the drone RCs transmit signals in the 2.4 GHz band. There are 17 drone RCs from eight different manufacturers and ~1000 RF signals per drone RC, each spanning a duration of 0.25 ms.

2021 IEEE Aerospace Conference (50100), 2021

We propose a security architecture that uses the principle of chaos for UAV secure communication.... more We propose a security architecture that uses the principle of chaos for UAV secure communication. A UAV, identified as an aerial base station (ABS), communicates with a ground base station (GBS) over a wireless radio frequency (RF) channel. The communication units of the ABS and GBS have dynamics according to the logistic map. The map is chaotic in the appropriate parameter space. Its states are non-periodic, broadband, and noise-like in the frequency domain. They

arXiv: Signal Processing, 2019

This paper investigates the problem of detection and classification of unmanned aerial vehicles (... more 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 naive 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 lea...

arXiv: Signal Processing, 2018

The future 5G networks are expected to use millimeter wave (mmWave) frequency bands, mainly due t... more The future 5G networks are expected to use millimeter wave (mmWave) frequency bands, mainly due to the availability of large unused spectrum. However, due to high path loss at mmWave frequencies, coverage of mmWave signals can get severely reduced, especially for non-line-of-sight (NLOS) scenarios. In this work, we study the use of passive metallic reflectors of different shapes/sizes to improve mmWave signal coverage for indoor NLOS scenarios. Software defined radio based mmWave transceiver platforms operating at 28 GHz are used for indoor measurements. Subsequently, ray tracing (RT) simulations are carried out in a similar environment using Remcom Wireless InSite software. The cumulative distribution functions of the received signal strength for the RT simulations in the area of interest are observed to be reasonably close with those obtained from the measurements. Our measurements and RT simulations both show that there is significant (on the order of 20 dB) power gain obtained w...

2021 IEEE 32nd Annual International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), 2021