Pub Date : 2021-11-29DOI: 10.1109/MILCOM52596.2021.9652977
Suzanna Lamar, J. Gosselin, Ivan Caceres, Sarah Kapple, A. Jayasumana
Making use of spectrally diverse communications links to re-route traffic in response to dynamic environments to manage network bottlenecks has become essential in order to guarantee message delivery across heterogeneous networks. We propose an innovative, proactive Congestion Aware Intent-Based Routing (CONAIR) architecture that can select among available communication link resources based on quality of service (QoS) metrics to support continuous information exchange between networked participants. The CONAIR architecture utilizes a Network Controller (NC) and artificial intelligence (AI) to re-route traffic based on traffic priority, fundamental to increasing end user quality of experience (QoE) and mission effectiveness. The CONAIR architecture provides network behavior prediction, and can mitigate congestion prior to its occurrence unlike traditional static routing techniques, e.g. Open Shortest Path First (OSPF), which are prone to congestion due to infrequent routing table updates. Modeling and simulation (M&S) was performed on a multi-hop network in order to characterize the resiliency and scalability benefits of CONAIR over OSPF routing-based frameworks. Results demonstrate that for varying traffic profiles, packet loss and end-to-end latency is minimized.
{"title":"Congestion Aware Intent-Based Routing using Graph Neural Networks for Improved Quality of Experience in Heterogeneous Networks","authors":"Suzanna Lamar, J. Gosselin, Ivan Caceres, Sarah Kapple, A. Jayasumana","doi":"10.1109/MILCOM52596.2021.9652977","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9652977","url":null,"abstract":"Making use of spectrally diverse communications links to re-route traffic in response to dynamic environments to manage network bottlenecks has become essential in order to guarantee message delivery across heterogeneous networks. We propose an innovative, proactive Congestion Aware Intent-Based Routing (CONAIR) architecture that can select among available communication link resources based on quality of service (QoS) metrics to support continuous information exchange between networked participants. The CONAIR architecture utilizes a Network Controller (NC) and artificial intelligence (AI) to re-route traffic based on traffic priority, fundamental to increasing end user quality of experience (QoE) and mission effectiveness. The CONAIR architecture provides network behavior prediction, and can mitigate congestion prior to its occurrence unlike traditional static routing techniques, e.g. Open Shortest Path First (OSPF), which are prone to congestion due to infrequent routing table updates. Modeling and simulation (M&S) was performed on a multi-hop network in order to characterize the resiliency and scalability benefits of CONAIR over OSPF routing-based frameworks. Results demonstrate that for varying traffic profiles, packet loss and end-to-end latency is minimized.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121115491","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-11-29DOI: 10.1109/MILCOM52596.2021.9652967
Tarak Arbi, B. Geller, O. Pasquero
Jamming attacks can severely limit wireless networks availability and can cause serious damage, in particular for tactical applications. Over the past decades, Direct-Sequence Spread Spectrum (DSSS) has been used to enhance resistance to jamming. In this paper, we first analyze the performance of the DSSS modulation in the presence of malicious jamming; we take into account by considering different physical phenomena such as a large Doppler shift and we use at the receiver side robust synchronization algorithms. We then propose to consider jointly rotated constellations and the DSSS technique in order to enhance robustness against jamming, while keeping reasonable complexity. Simulations results underline the good performance of our proposal as it shows a gain of several dBs compared to the DSSS technique with conventional constellations.
{"title":"Direct-Sequence Spread Spectrum with Signal Space Diversity for High Resistance to Jamming","authors":"Tarak Arbi, B. Geller, O. Pasquero","doi":"10.1109/MILCOM52596.2021.9652967","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9652967","url":null,"abstract":"Jamming attacks can severely limit wireless networks availability and can cause serious damage, in particular for tactical applications. Over the past decades, Direct-Sequence Spread Spectrum (DSSS) has been used to enhance resistance to jamming. In this paper, we first analyze the performance of the DSSS modulation in the presence of malicious jamming; we take into account by considering different physical phenomena such as a large Doppler shift and we use at the receiver side robust synchronization algorithms. We then propose to consider jointly rotated constellations and the DSSS technique in order to enhance robustness against jamming, while keeping reasonable complexity. Simulations results underline the good performance of our proposal as it shows a gain of several dBs compared to the DSSS technique with conventional constellations.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"334 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116529279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-11-29DOI: 10.1109/MILCOM52596.2021.9653046
Xavier Leturc, Elie Janin, C. Martret
In this paper, we present results coming from the Multi bAnd Efficient Networks for Ad hoc communications (MAENA) project. We present a distributed dynamic channel assignment (DDCA) algorithm for military clustered ad hoc networks that has been designed during the project. To design the proposed algorithm, we start from an existing algorithm called greedy based dynamic channel assignment (GBDCA). We first identify the main drawbacks of this algorithm, and then we propose modifications to alleviate these drawbacks yielding the proposed GBDCA++ solution. We also explain how to implement the GBDCA++ algorithm in a military waveform developed in the MAENA project. Finally, we show the superiority of GBDCA++ over GBDCA by providing high fidelity simulation results obtained using the simulator developed within the framework of the MAENA project.
{"title":"Distributed Dynamic Channel Assignment in Military Ad Hoc Networks within the MAENA Project: New Algorithm and High Fidelity Simulation Results","authors":"Xavier Leturc, Elie Janin, C. Martret","doi":"10.1109/MILCOM52596.2021.9653046","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9653046","url":null,"abstract":"In this paper, we present results coming from the Multi bAnd Efficient Networks for Ad hoc communications (MAENA) project. We present a distributed dynamic channel assignment (DDCA) algorithm for military clustered ad hoc networks that has been designed during the project. To design the proposed algorithm, we start from an existing algorithm called greedy based dynamic channel assignment (GBDCA). We first identify the main drawbacks of this algorithm, and then we propose modifications to alleviate these drawbacks yielding the proposed GBDCA++ solution. We also explain how to implement the GBDCA++ algorithm in a military waveform developed in the MAENA project. Finally, we show the superiority of GBDCA++ over GBDCA by providing high fidelity simulation results obtained using the simulator developed within the framework of the MAENA project.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116939652","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-11-29DOI: 10.1109/MILCOM52596.2021.9653103
Bertram Schütz, Stefanie Thieme, Christoph Fuchs, Daniel Weber, N. Aschenbruck
This paper formalizes and evaluates a promising technique to overcome packet loss in tactical scenarios, called Network Coding-based Multi-Path Forward Erasure Correction (CoMPEC). Thereby, encoded redundancy packets are sent over a secondary path to correct packet loss on the main path without the usage of feedback or retransmissions. Formal equations are presented to calculate the benefits in terms of packet loss rate after decoding and coding gain. To evaluate the potential for tactical scenarios, a simulation was conducted, which is based on the Anglova path loss data. The presented evaluation verifies CoMPEC's ability to significantly reduce the packet loss rate at the receiver, if the scheme is applicable.
{"title":"Network Coding-based Multi-Path Forward Erasure Correction for Tactical Scenarios","authors":"Bertram Schütz, Stefanie Thieme, Christoph Fuchs, Daniel Weber, N. Aschenbruck","doi":"10.1109/MILCOM52596.2021.9653103","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9653103","url":null,"abstract":"This paper formalizes and evaluates a promising technique to overcome packet loss in tactical scenarios, called Network Coding-based Multi-Path Forward Erasure Correction (CoMPEC). Thereby, encoded redundancy packets are sent over a secondary path to correct packet loss on the main path without the usage of feedback or retransmissions. Formal equations are presented to calculate the benefits in terms of packet loss rate after decoding and coding gain. To evaluate the potential for tactical scenarios, a simulation was conducted, which is based on the Anglova path loss data. The presented evaluation verifies CoMPEC's ability to significantly reduce the packet loss rate at the receiver, if the scheme is applicable.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"81 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122083182","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-11-29DOI: 10.1109/MILCOM52596.2021.9652921
J. Sigholm, Emil Larsson
In this paper we revisit a study presented at MILCOM 2014. Our goal then was to determine the utility of implanting a vulnerability into a cybersecurity software protocol to an actor planning to execute an offensive cyber operation. Based on a case study describing the then recently discovered Heartbleed bug as an offensive cyber operation, a model was devised to estimate the adoption rate of an implanted flaw in OpenSSL. Using the adoption rate of the cryptographic protocol Transport Layer Security version 1.2 as a proxy, we predicted that the global adoption of the vulnerability of at least 50% would take approximately three years, while surpassing 75% adoption would take four years. Compared to subsequently collected real-world data, these forecasts turned out to be surprisingly accurate. An evaluation of our proposed model shows that it yields results with a root-mean-square error of only 1.2% over the forecasting period. Thus, it has a significant degree of predictive power. Although the model may not be generalizable to describe the adoption of any software protocol, the finding helps validate our previously drawn conclusion that exploiting implanted cyber vulnerabilities, in a scenario like the one presented, requires a planning horizon of multiple years. However, as society becomes further dependent on the cyber domain, the utility of intentional vulnerability implantation is likely an exercise in diminishing returns. For a defender, however, our model development process could be useful to forecast the time required for flawed protocols to be phased out.
{"title":"Cyber Vulnerability Implantation Revisited","authors":"J. Sigholm, Emil Larsson","doi":"10.1109/MILCOM52596.2021.9652921","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9652921","url":null,"abstract":"In this paper we revisit a study presented at MILCOM 2014. Our goal then was to determine the utility of implanting a vulnerability into a cybersecurity software protocol to an actor planning to execute an offensive cyber operation. Based on a case study describing the then recently discovered Heartbleed bug as an offensive cyber operation, a model was devised to estimate the adoption rate of an implanted flaw in OpenSSL. Using the adoption rate of the cryptographic protocol Transport Layer Security version 1.2 as a proxy, we predicted that the global adoption of the vulnerability of at least 50% would take approximately three years, while surpassing 75% adoption would take four years. Compared to subsequently collected real-world data, these forecasts turned out to be surprisingly accurate. An evaluation of our proposed model shows that it yields results with a root-mean-square error of only 1.2% over the forecasting period. Thus, it has a significant degree of predictive power. Although the model may not be generalizable to describe the adoption of any software protocol, the finding helps validate our previously drawn conclusion that exploiting implanted cyber vulnerabilities, in a scenario like the one presented, requires a planning horizon of multiple years. However, as society becomes further dependent on the cyber domain, the utility of intentional vulnerability implantation is likely an exercise in diminishing returns. For a defender, however, our model development process could be useful to forecast the time required for flawed protocols to be phased out.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124192484","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-11-29DOI: 10.1109/MILCOM52596.2021.9653050
Mattia Fogli, Geert L. J. Pingen, Thomas Kudla, S. Webb, Niranjan Suri, H. Bastiaansen
Nowadays, ever-increasing processing and storage resources are available at all echelons, from operations centers to tactical units. However, tactical-edge communications still suffer from scarce network resources such as limited bandwidth, intermittent connectivity, and variable latency. In addition, modern military missions typically involve coalition operations, where heterogeneous mission partners (even belonging to different nations) cooperate in the field. As a result, the distribution of mission-critical information is more complicated than ever. On the one hand, the dynamic nature of the tactical environment frequently disrupts communications. On the other hand, individual resource-sharing policies prevent mission partners from taking full advantage of the available resources in situ. The NATO IST-168 RTG has been exploring commercial-off-the-shelf orchestration technologies for implementing a federated cloud architecture that enables adaptive information processing and dissemination while living within the constraints of the tactical domain. This paper is a follow-up study that assesses the behaviour of Kubernetes under the disadvantaged network conditions characterizing tactical edge networks.
{"title":"Towards a COTS-Enabled Federated Cloud Architecture for Adaptive C2 in Coalition Tactical Operations: A Performance Analysis of Kubernetes","authors":"Mattia Fogli, Geert L. J. Pingen, Thomas Kudla, S. Webb, Niranjan Suri, H. Bastiaansen","doi":"10.1109/MILCOM52596.2021.9653050","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9653050","url":null,"abstract":"Nowadays, ever-increasing processing and storage resources are available at all echelons, from operations centers to tactical units. However, tactical-edge communications still suffer from scarce network resources such as limited bandwidth, intermittent connectivity, and variable latency. In addition, modern military missions typically involve coalition operations, where heterogeneous mission partners (even belonging to different nations) cooperate in the field. As a result, the distribution of mission-critical information is more complicated than ever. On the one hand, the dynamic nature of the tactical environment frequently disrupts communications. On the other hand, individual resource-sharing policies prevent mission partners from taking full advantage of the available resources in situ. The NATO IST-168 RTG has been exploring commercial-off-the-shelf orchestration technologies for implementing a federated cloud architecture that enables adaptive information processing and dissemination while living within the constraints of the tactical domain. This paper is a follow-up study that assesses the behaviour of Kubernetes under the disadvantaged network conditions characterizing tactical edge networks.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"76 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122389480","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-11-29DOI: 10.1109/MILCOM52596.2021.9652981
Farah Arabian, M. Rice
This paper examines the relationship between polarization diversity combining and equalization in the context of aeronautical mobile telemetry (AMT). The receive antennas currently used in AMT combine the linear polarizations using a 90° hybrid coupler to synthesize circularly polarized signals. Maximum likelihood (ML) polarization combining is developed for this application. Computer simulation results show that equalizing an ML combined channel achieves a lower bit error rate than equalizing circularly combined signals. The performance improvement is achieved at the cost of estimating the channel.
{"title":"Polarization Combining and Equalization for Aeronautical Mobile Telemetry","authors":"Farah Arabian, M. Rice","doi":"10.1109/MILCOM52596.2021.9652981","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9652981","url":null,"abstract":"This paper examines the relationship between polarization diversity combining and equalization in the context of aeronautical mobile telemetry (AMT). The receive antennas currently used in AMT combine the linear polarizations using a 90° hybrid coupler to synthesize circularly polarized signals. Maximum likelihood (ML) polarization combining is developed for this application. Computer simulation results show that equalizing an ML combined channel achieves a lower bit error rate than equalizing circularly combined signals. The performance improvement is achieved at the cost of estimating the channel.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130734873","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-11-29DOI: 10.1109/MILCOM52596.2021.9652916
S. Venkatesan, Harshvardhan Digvijay Sikka, R. Izmailov, R. Chadha, Alina Oprea, Michael J. de Lucia
Among many application domains of machine learning in real-world settings, cyber security can benefit from more automated techniques to combat sophisticated adversaries. Modern network intrusion detection systems leverage machine learning models on network logs to proactively detect cyber attacks. However, the risk of adversarial attacks against machine learning used in these cyber settings is not fully explored. In this paper, we investigate poisoning attacks at training time against machine learning models in constrained cyber environments such as network intrusion detection; we also explore mitigations of such attacks based on training data sanitization. We consider the setting of poisoning availability attacks, in which an attacker can insert a set of poisoned samples at training time with the goal of degrading the accuracy of the deployed model. We design a white-box, realizable poisoning attack that reduced the original model accuracy from 95% to less than 50 % by generating mislabeled samples in close vicinity of a selected subset of training points. We also propose a novel Nested Training method as a defense against these attacks. Our defense includes a diversified ensemble of classifiers, each trained on a different subset of the training set. We use the disagreement of the classifiers' predictions as a data sanitization method, and show that an ensemble of 10 SVM classifiers is resilient to a large fraction of poisoning samples, up to 30% of the training data.
{"title":"Poisoning Attacks and Data Sanitization Mitigations for Machine Learning Models in Network Intrusion Detection Systems","authors":"S. Venkatesan, Harshvardhan Digvijay Sikka, R. Izmailov, R. Chadha, Alina Oprea, Michael J. de Lucia","doi":"10.1109/MILCOM52596.2021.9652916","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9652916","url":null,"abstract":"Among many application domains of machine learning in real-world settings, cyber security can benefit from more automated techniques to combat sophisticated adversaries. Modern network intrusion detection systems leverage machine learning models on network logs to proactively detect cyber attacks. However, the risk of adversarial attacks against machine learning used in these cyber settings is not fully explored. In this paper, we investigate poisoning attacks at training time against machine learning models in constrained cyber environments such as network intrusion detection; we also explore mitigations of such attacks based on training data sanitization. We consider the setting of poisoning availability attacks, in which an attacker can insert a set of poisoned samples at training time with the goal of degrading the accuracy of the deployed model. We design a white-box, realizable poisoning attack that reduced the original model accuracy from 95% to less than 50 % by generating mislabeled samples in close vicinity of a selected subset of training points. We also propose a novel Nested Training method as a defense against these attacks. Our defense includes a diversified ensemble of classifiers, each trained on a different subset of the training set. We use the disagreement of the classifiers' predictions as a data sanitization method, and show that an ensemble of 10 SVM classifiers is resilient to a large fraction of poisoning samples, up to 30% of the training data.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129864840","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-11-29DOI: 10.1109/MILCOM52596.2021.9653072
Wenhan Zhang, M. Krunz, G. Ditzler
Deep neural networks (DNNs) have recently been applied in the classification of radio frequency (RF) signals. One use case of interest relates to the discernment between different wireless technologies that share the spectrum. Although highly accurate DNN classifiers have been proposed, preliminary research points to the vulnerability of these classifiers to adversarial machine learning (AML) attacks. In one such attack, a surrogate DNN model is trained by the attacker to produce intelligently crafted low-power “perturbations” that degrade the classification accuracy of the legitimate classifier. In this paper, we design four DNN-based classifiers for the identification of Wi-Fi, 5G NR-Unlicensed (NR-U), and LTE LAA transmissions over the 5 GHz UNII bands. Our DNN models include both convolutional neural networks (CNNs) as well as several recurrent neural networks (RNNs) models, particularly LSTM and Bidirectional LSTM (BiLSTM) networks. We demonstrate the high classification accuracy of these models under “benign” (non-adversarial) noise. We then study the efficacy of these classifiers under AML-based perturbations. Specifically, we use the fast gradient sign method (FGSM) to generate adversarial perturbations. Different attack scenarios are studied, depending on how much information the attacker has about the defender's classifier. In one extreme scenario, called “white-box” attack, the attacker has full knowledge of the defender's DNN, including its hyperparameters, its training dataset, and even the seeds used to train the network. This attack is shown to significantly degrade the classification accuracy even when the FGSM-based perturbations are low power, i.e., the received SNR is relatively high. We then consider more realistic attack scenarios, where the attacker has partial or no knowledge of the defender's classifier. Even under limited knowledge, adversarial perturbations can still lead to significant reduction in the classification accuracy, relative to classification under AWGN with the same SNR level.
{"title":"Intelligent Jamming of Deep Neural Network Based Signal Classification for Shared Spectrum","authors":"Wenhan Zhang, M. Krunz, G. Ditzler","doi":"10.1109/MILCOM52596.2021.9653072","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9653072","url":null,"abstract":"Deep neural networks (DNNs) have recently been applied in the classification of radio frequency (RF) signals. One use case of interest relates to the discernment between different wireless technologies that share the spectrum. Although highly accurate DNN classifiers have been proposed, preliminary research points to the vulnerability of these classifiers to adversarial machine learning (AML) attacks. In one such attack, a surrogate DNN model is trained by the attacker to produce intelligently crafted low-power “perturbations” that degrade the classification accuracy of the legitimate classifier. In this paper, we design four DNN-based classifiers for the identification of Wi-Fi, 5G NR-Unlicensed (NR-U), and LTE LAA transmissions over the 5 GHz UNII bands. Our DNN models include both convolutional neural networks (CNNs) as well as several recurrent neural networks (RNNs) models, particularly LSTM and Bidirectional LSTM (BiLSTM) networks. We demonstrate the high classification accuracy of these models under “benign” (non-adversarial) noise. We then study the efficacy of these classifiers under AML-based perturbations. Specifically, we use the fast gradient sign method (FGSM) to generate adversarial perturbations. Different attack scenarios are studied, depending on how much information the attacker has about the defender's classifier. In one extreme scenario, called “white-box” attack, the attacker has full knowledge of the defender's DNN, including its hyperparameters, its training dataset, and even the seeds used to train the network. This attack is shown to significantly degrade the classification accuracy even when the FGSM-based perturbations are low power, i.e., the received SNR is relatively high. We then consider more realistic attack scenarios, where the attacker has partial or no knowledge of the defender's classifier. Even under limited knowledge, adversarial perturbations can still lead to significant reduction in the classification accuracy, relative to classification under AWGN with the same SNR level.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128193018","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-11-29DOI: 10.1109/MILCOM52596.2021.9653128
Kyle Willstatter, M. Zoltowski
The high PAPR of OFDM transmission leads to power/cost inefficiencies in amplifier use and/or spectral noise from clipping effects. To avoid these issues, we propose transmitting a complementary sequence pair whose aperiodic autocorrelations sum to a delta function in such a way that the amplitude of the signal is constant. This enables the use of low-cost nonlinear amplifiers operating at full power. The sequence pair is constructed iteratively, by sequential encoding of information symbols onto the pair such that the sequences remain complementary. The structure of these sequences and the resulting constant-envelope signal are analyzed, leading to methods of symbol extraction and the results of a decoding error. Finally, we extend the discussion to two dimensional sequence pairs, for use in mmWave/MIMO systems where the inefficiencies of a high PAPR are even more acute.
{"title":"Complementary Sequence Construction for Constant-Envelope OFDM Transmission Enabling Nonlinear Amplification and Clipping","authors":"Kyle Willstatter, M. Zoltowski","doi":"10.1109/MILCOM52596.2021.9653128","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9653128","url":null,"abstract":"The high PAPR of OFDM transmission leads to power/cost inefficiencies in amplifier use and/or spectral noise from clipping effects. To avoid these issues, we propose transmitting a complementary sequence pair whose aperiodic autocorrelations sum to a delta function in such a way that the amplitude of the signal is constant. This enables the use of low-cost nonlinear amplifiers operating at full power. The sequence pair is constructed iteratively, by sequential encoding of information symbols onto the pair such that the sequences remain complementary. The structure of these sequences and the resulting constant-envelope signal are analyzed, leading to methods of symbol extraction and the results of a decoding error. Finally, we extend the discussion to two dimensional sequence pairs, for use in mmWave/MIMO systems where the inefficiencies of a high PAPR are even more acute.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125906007","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：