Pub Date : 2021-11-29DOI: 10.1109/MILCOM52596.2021.9652985
David C. Last, M. Atighetchi, P. Pal, Ryan Toner
The US military is focused on transitioning its warf-ighting philosophy towards the Multi-Domain Command and Control (MDC2) concept, which integrates lands, sea, air, space, and cyberspace forces into a unified planning and execution structure. This structure depends on reliable exchange of plans and intelligence reports, which is hindered by the development of adversary capabilities in the areas of radio jamming and other communications-denying technologies. The battlefield of the future will consist of dispersed units with intermittent, constrained communications who nevertheless need to maintain a consistent view of shared plans and information. We are developing a proof-of-concept prototype MDC2 system to explore the requirements and limitations of a solution in this space. There are many distributed database solutions that could be used for this prototype; upon investigation, we believe that Distributed Ledger Technologies (DLT) are an ideal fit for the particular constraints of this use case. We have selected Hyperledger Fabric as the particular DLT implementation on which we are building this prototype. As part of this research, we are developing a Hyperledger-based prototype as well as a baseline implementation that reflects how the DoD handles this process today. We will use both of these systems in a series of experiments that evaluate how well they perform in the areas of database consistency, constrained tactical networks, and conditional authorities, and we will measure their performance using a suite of metrics that will allow us to compare the DLT-based prototype with the baseline system. This paper outlines the challenges we will address through this research; the completed research will encompass our solutions to these challenges.
{"title":"Using Distributed Ledgers For Command and Control – Concepts and Challenges","authors":"David C. Last, M. Atighetchi, P. Pal, Ryan Toner","doi":"10.1109/MILCOM52596.2021.9652985","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9652985","url":null,"abstract":"The US military is focused on transitioning its warf-ighting philosophy towards the Multi-Domain Command and Control (MDC2) concept, which integrates lands, sea, air, space, and cyberspace forces into a unified planning and execution structure. This structure depends on reliable exchange of plans and intelligence reports, which is hindered by the development of adversary capabilities in the areas of radio jamming and other communications-denying technologies. The battlefield of the future will consist of dispersed units with intermittent, constrained communications who nevertheless need to maintain a consistent view of shared plans and information. We are developing a proof-of-concept prototype MDC2 system to explore the requirements and limitations of a solution in this space. There are many distributed database solutions that could be used for this prototype; upon investigation, we believe that Distributed Ledger Technologies (DLT) are an ideal fit for the particular constraints of this use case. We have selected Hyperledger Fabric as the particular DLT implementation on which we are building this prototype. As part of this research, we are developing a Hyperledger-based prototype as well as a baseline implementation that reflects how the DoD handles this process today. We will use both of these systems in a series of experiments that evaluate how well they perform in the areas of database consistency, constrained tactical networks, and conditional authorities, and we will measure their performance using a suite of metrics that will allow us to compare the DLT-based prototype with the baseline system. This paper outlines the challenges we will address through this research; the completed research will encompass our solutions to these challenges.","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":"114471729","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/MILCOM52596.2021.9652935
J. Norris, J. Nieto
Communications Security is Achieved in the Physical Layer with the utilization of Low Probability of Detection (LPD) and Low Probability of Intercept (LPI) Modulation Techniques. Standard modulation techniques have cyclical time varying properties that can be exploited by signals intelligence (SIGINT) for signal detection, traffic analysis, and signal interception. There is a large body of analysis on wide-sense cyclostationary features of these modulation techniques with the conclusion that the primary cause of periodic cyclostationary features is the modulation symbol rate. Deterministic variability in the symbol rate can be used to reduce or eliminate cyclostationary features. Several techniques have been proposed, including DAC clock dithering. In this paper we provide a description of a polyphase resampler application to provide symbol rate dithering and an analysis of the dithering on the second, and higher order cyclostationary features. Multiple modulation types are included in the analysis. ADC de-dithering is typically not an option at the receiver - the delays inherent in standard receiver architectures prohibit real-time control of the ADC during signal demodulation while the polyphase resampler can be applied digitally within the receiver at matched filter sample rates.
{"title":"Application of Sub-Sample Dithering to Reduce Probability of Signal Detection","authors":"J. Norris, J. Nieto","doi":"10.1109/MILCOM52596.2021.9652935","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9652935","url":null,"abstract":"Communications Security is Achieved in the Physical Layer with the utilization of Low Probability of Detection (LPD) and Low Probability of Intercept (LPI) Modulation Techniques. Standard modulation techniques have cyclical time varying properties that can be exploited by signals intelligence (SIGINT) for signal detection, traffic analysis, and signal interception. There is a large body of analysis on wide-sense cyclostationary features of these modulation techniques with the conclusion that the primary cause of periodic cyclostationary features is the modulation symbol rate. Deterministic variability in the symbol rate can be used to reduce or eliminate cyclostationary features. Several techniques have been proposed, including DAC clock dithering. In this paper we provide a description of a polyphase resampler application to provide symbol rate dithering and an analysis of the dithering on the second, and higher order cyclostationary features. Multiple modulation types are included in the analysis. ADC de-dithering is typically not an option at the receiver - the delays inherent in standard receiver architectures prohibit real-time control of the ADC during signal demodulation while the polyphase resampler can be applied digitally within the receiver at matched filter sample rates.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"79 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":"121620856","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/MILCOM52596.2021.9652937
A. Poylisher, A. Cichocki, K. Guo, J. Hunziker, L. Kant, B. Krishnamachari, A. Avestimehr, M. Annavaram
We present Tactical Jupiter, an adaptation of the recently developed Jupiter framework for scheduling of dispersed computations on heterogeneous resources to tactical MANETs. Tactical Jupiter addresses the challenges to distributed scheduling posed by intermittent connectivity and scarce/variable bandwidth, variable computational resource utilization by background load, and node attrition. Our key contributions include: (a) disruption handling via increased autonomy of task executors, (b) low-overhead ML-based task completion time estimation in presence of background load, and (c) resilient dissemination mechanisms for monitoring information.
{"title":"Tactical Jupiter: Dynamic Scheduling of Dispersed Computations in Tactical MANETs","authors":"A. Poylisher, A. Cichocki, K. Guo, J. Hunziker, L. Kant, B. Krishnamachari, A. Avestimehr, M. Annavaram","doi":"10.1109/MILCOM52596.2021.9652937","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9652937","url":null,"abstract":"We present Tactical Jupiter, an adaptation of the recently developed Jupiter framework for scheduling of dispersed computations on heterogeneous resources to tactical MANETs. Tactical Jupiter addresses the challenges to distributed scheduling posed by intermittent connectivity and scarce/variable bandwidth, variable computational resource utilization by background load, and node attrition. Our key contributions include: (a) disruption handling via increased autonomy of task executors, (b) low-overhead ML-based task completion time estimation in presence of background load, and (c) resilient dissemination mechanisms for monitoring information.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"183 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":"123295240","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/MILCOM52596.2021.9653093
Manas Pradhan, Marco Manso, James R. Michaelis
The evolving fields of Information and Communications Technology (ICT) and the Internet of Things (IoT) have allowed a multitude of state and non-state actors to leverage remote, ubiquitous and context rich computing for solving future challenges. Individually, standards underlying ICT/IoT serve the purpose they were devised for in supporting predetermined use cases and application domains, but often present significant challenges when underlying technologies must be re-purposed or integrated in new ways. Corresponding technology integration challenges are often addressed through ad-hoc solutions (e.g., stove-piped interfaces, data federation and exchange services) which make it very difficult to scale and adapt to the needs of cross-coalition Command Control (C2) infrastructures. Following from NATO IST-176 efforts, this paper reviews key design requirements for developing NATO C2 Data Models to facilitate COTS IoT and NATO STANAG interoperability. Here, the idea is not to propose an entirely new standard but to identify and catalogue methods to reuse existing standards to improve interoperability as well as introduce extensibility to enable next-generation coalition operations.
{"title":"Concepts and Directions for Future IoT and C2 Interoperability","authors":"Manas Pradhan, Marco Manso, James R. Michaelis","doi":"10.1109/MILCOM52596.2021.9653093","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9653093","url":null,"abstract":"The evolving fields of Information and Communications Technology (ICT) and the Internet of Things (IoT) have allowed a multitude of state and non-state actors to leverage remote, ubiquitous and context rich computing for solving future challenges. Individually, standards underlying ICT/IoT serve the purpose they were devised for in supporting predetermined use cases and application domains, but often present significant challenges when underlying technologies must be re-purposed or integrated in new ways. Corresponding technology integration challenges are often addressed through ad-hoc solutions (e.g., stove-piped interfaces, data federation and exchange services) which make it very difficult to scale and adapt to the needs of cross-coalition Command Control (C2) infrastructures. Following from NATO IST-176 efforts, this paper reviews key design requirements for developing NATO C2 Data Models to facilitate COTS IoT and NATO STANAG interoperability. Here, the idea is not to propose an entirely new standard but to identify and catalogue methods to reuse existing standards to improve interoperability as well as introduce extensibility to enable next-generation coalition operations.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"42 8 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":"129881363","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/MILCOM52596.2021.9652882
Gautam Trivedi, B. Jabbari
Wireless multi-hop networks are rapidly becoming an integral part of next generation mobile communications. These networks are highly scalable, self-organizing, dynamic, and may share spectrum with others while operating in resource-constrained environments. As a result, these networks are highly susceptible to link failure. In particular, wireless multi-hop sensor networks are required to maintain a high level of resiliency in order to deliver sensor data with minimum latency. One way to achieve such resiliency is by maximizing the probability of maintaining a high level of topological connectivity. In this paper, we develop a graph-theoretic model that maintains a high level of connectivity in the presence of internal and external interference. Given a pool of available channels, we utilize spectrum sensing to determine the state of each channel, achieve channel state consensus across the network using distributed average consensus approach and use graph coloring for channel allocation in order to minimize interference experienced from adjacent nodes as well as external interference sources. These models, corroborated by simulation, are then utilized to quantify the improvement in network resiliency as measured by link connectivity.
{"title":"On Wireless Link Connectivity for Resilient Multi-Hop Networks","authors":"Gautam Trivedi, B. Jabbari","doi":"10.1109/MILCOM52596.2021.9652882","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9652882","url":null,"abstract":"Wireless multi-hop networks are rapidly becoming an integral part of next generation mobile communications. These networks are highly scalable, self-organizing, dynamic, and may share spectrum with others while operating in resource-constrained environments. As a result, these networks are highly susceptible to link failure. In particular, wireless multi-hop sensor networks are required to maintain a high level of resiliency in order to deliver sensor data with minimum latency. One way to achieve such resiliency is by maximizing the probability of maintaining a high level of topological connectivity. In this paper, we develop a graph-theoretic model that maintains a high level of connectivity in the presence of internal and external interference. Given a pool of available channels, we utilize spectrum sensing to determine the state of each channel, achieve channel state consensus across the network using distributed average consensus approach and use graph coloring for channel allocation in order to minimize interference experienced from adjacent nodes as well as external interference sources. These models, corroborated by simulation, are then utilized to quantify the improvement in network resiliency as measured by link connectivity.","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":"130216284","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/MILCOM52596.2021.9653015.
Yuxiao Li, S. Mazuelas, Yuan Shen
Ultra-wideband (UWB)-based techniques, while becoming mainstream approaches for high-accurate positioning, tend to be challenged by ranging bias in harsh environments. The emerging learning-based methods for error mitigation have shown great performance improvement via exploiting high semantic features from raw data. However, these methods rely heavily on fully labeled data, leading to a high cost for data acquisition. We present a learning framework based on weak supervision for UWB ranging error mitigation. Specifically, we propose a deep learning method based on the generalized expectation-maximization (GEM) algorithm for robust UWB ranging error mitigation under weak supervision. Such method integrate probabilistic modeling into the deep learning scheme, and adopt weakly supervised labels as prior information. Extensive experiments in various supervision scenarios illustrate the superiority of the proposed method.
{"title":"Deep GEM-Based Network for Weakly Supervised UWB Ranging Error Mitigation","authors":"Yuxiao Li, S. Mazuelas, Yuan Shen","doi":"10.1109/MILCOM52596.2021.9653015.","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9653015.","url":null,"abstract":"Ultra-wideband (UWB)-based techniques, while becoming mainstream approaches for high-accurate positioning, tend to be challenged by ranging bias in harsh environments. The emerging learning-based methods for error mitigation have shown great performance improvement via exploiting high semantic features from raw data. However, these methods rely heavily on fully labeled data, leading to a high cost for data acquisition. We present a learning framework based on weak supervision for UWB ranging error mitigation. Specifically, we propose a deep learning method based on the generalized expectation-maximization (GEM) algorithm for robust UWB ranging error mitigation under weak supervision. Such method integrate probabilistic modeling into the deep learning scheme, and adopt weakly supervised labels as prior information. Extensive experiments in various supervision scenarios illustrate the superiority of the proposed method.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"21 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":"129983255","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/MILCOM52596.2021.9652997
Anne M. Tall, C. Zou, Jun Wang
This paper provides an overview of the security service controls that are applied in a big data processing (BDP) system to defend against cyber security attacks. We validate this approach by modeling attacks and effectiveness of security service controls in a sequence of states and transitions. This Finite State Machine (FSM) approach uses the probable effectiveness of security service controls, as defined in the National Institute of Standards and Technology (NIST) Risk Management Framework (RMF). The attacks used in the model are defined in the ATT&CK™ framework. Five different BDP security architecture configurations are considered, spanning from a low-cost default BDP configuration to a more expensive, industry supported layered security architecture. The analysis demonstrates the importance of a multi-layer approach to implementing security in BDP systems. With increasing interest in using BDP systems to analyze sensitive data sets, it is important to understand and justify BDP security architecture configurations with their significant costs. The output of the model demonstrates that over the run time, larger investment in security service controls results in significantly more uptime. There is a significant increase in uptime with a linear increase in security service control investment. We believe that these results support our recommended BDP security architecture. That is, a layered architecture with security service controls integrated into the user interface, boundary, central management of security policies, and applications that incorporate privacy preserving programs. These results enable making BDP systems operational for sensitive data accessed in a multi-tenant environment.
{"title":"Integrating Cybersecurity Into a Big Data Ecosystem","authors":"Anne M. Tall, C. Zou, Jun Wang","doi":"10.1109/MILCOM52596.2021.9652997","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9652997","url":null,"abstract":"This paper provides an overview of the security service controls that are applied in a big data processing (BDP) system to defend against cyber security attacks. We validate this approach by modeling attacks and effectiveness of security service controls in a sequence of states and transitions. This Finite State Machine (FSM) approach uses the probable effectiveness of security service controls, as defined in the National Institute of Standards and Technology (NIST) Risk Management Framework (RMF). The attacks used in the model are defined in the ATT&CK™ framework. Five different BDP security architecture configurations are considered, spanning from a low-cost default BDP configuration to a more expensive, industry supported layered security architecture. The analysis demonstrates the importance of a multi-layer approach to implementing security in BDP systems. With increasing interest in using BDP systems to analyze sensitive data sets, it is important to understand and justify BDP security architecture configurations with their significant costs. The output of the model demonstrates that over the run time, larger investment in security service controls results in significantly more uptime. There is a significant increase in uptime with a linear increase in security service control investment. We believe that these results support our recommended BDP security architecture. That is, a layered architecture with security service controls integrated into the user interface, boundary, central management of security policies, and applications that incorporate privacy preserving programs. These results enable making BDP systems operational for sensitive data accessed in a multi-tenant environment.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"12 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":"124488235","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/MILCOM52596.2021.9653008
D. Breton, C. Haedrich
The fundamental challenge of complex (typically mountainous or urban) terrain for radio-frequency systems is that a direct line-of-sight is difficult or impossible to achieve between a radio transmitter and receiver. In the context of ground-based military operations, occupying line-of-sight positions to optimize communications or surveillance capabilities often exposes Soldiers and/or equipment to unacceptable risks from both conventional and electronic warfare. Propagation models for tactical use commonly analyze terrain only in the vertical plane containing both transmitter and receiver in order to simplify both the required data and computational burdens. However, these advantages come at the cost of ignoring reflections from topography within and outside the vertical plane path, which can have serious implications for radio-frequency direction finding, surveillance, and high-speed data transfer in complex terrain. This work summarizes our efforts to address these issues by developing a hybrid path loss model, one specifically designed for ground-to-ground radio links in complex rural terrain. The model uses an existing international-standard vertical plane diffraction model (VPM) to account for path losses associated with obstacles, and then augments those results with geospatially derived terrain reflection/scattering effects. Our prototype Terrain Scatter Augmented Vertical Plane Model (TSAVPM) provides physically credible path loss results in complex terrain at tactically relevant spatial scales (∼250 sq. km) and computational costs (under 40 seconds on a single 4.2 GHz central processing unit).
{"title":"Terrain-Scatter Augmented Vertical Plane Model for Radio Path Loss Estimation in Complex Terrain: (Invited Paper)","authors":"D. Breton, C. Haedrich","doi":"10.1109/MILCOM52596.2021.9653008","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9653008","url":null,"abstract":"The fundamental challenge of complex (typically mountainous or urban) terrain for radio-frequency systems is that a direct line-of-sight is difficult or impossible to achieve between a radio transmitter and receiver. In the context of ground-based military operations, occupying line-of-sight positions to optimize communications or surveillance capabilities often exposes Soldiers and/or equipment to unacceptable risks from both conventional and electronic warfare. Propagation models for tactical use commonly analyze terrain only in the vertical plane containing both transmitter and receiver in order to simplify both the required data and computational burdens. However, these advantages come at the cost of ignoring reflections from topography within and outside the vertical plane path, which can have serious implications for radio-frequency direction finding, surveillance, and high-speed data transfer in complex terrain. This work summarizes our efforts to address these issues by developing a hybrid path loss model, one specifically designed for ground-to-ground radio links in complex rural terrain. The model uses an existing international-standard vertical plane diffraction model (VPM) to account for path losses associated with obstacles, and then augments those results with geospatially derived terrain reflection/scattering effects. Our prototype Terrain Scatter Augmented Vertical Plane Model (TSAVPM) provides physically credible path loss results in complex terrain at tactically relevant spatial scales (∼250 sq. km) and computational costs (under 40 seconds on a single 4.2 GHz central processing unit).","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":"116724175","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/MILCOM52596.2021.9653101
Vikash Sehwag, J. W. Stokes, Cha Zhang
With the increasing adoption of deep learning in computer vision-based applications, it becomes critical to achieve robustness to real-world image transformations, such as geometric, photometric, and weather changes, even in the presence of an adversary. However, earlier work has focused on only a few transformations, such as image translation, rotation, or coloring. We close this gap by analyzing and improving robustness against twenty-four different physical transformations. First, we demonstrate that adversarial attacks based on each physical transformation significantly reduce the accuracy of deep neural networks. Next, we achieve robustness against these attacks based on adversarial training, where we show that single-step data augmentation significantly improves robustness against these attacks. We also demonstrate the generalization of robustness to these types of attacks, where robustness achieved against one attack also generalizes to some other attack vectors. Finally, we show that using an ensemble-based robust training approach, robustness against multiple attacks can be achieved simultaneously by a single network. In particular, our proposed method improves the aggregate robustness, against twenty-four different attacks, from 21.4% to 50.0% on the ImageNet dataset.
{"title":"Beyond $L_{p}$ Norms: Delving Deeper into Robustness to Physical Image Transformations","authors":"Vikash Sehwag, J. W. Stokes, Cha Zhang","doi":"10.1109/MILCOM52596.2021.9653101","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9653101","url":null,"abstract":"With the increasing adoption of deep learning in computer vision-based applications, it becomes critical to achieve robustness to real-world image transformations, such as geometric, photometric, and weather changes, even in the presence of an adversary. However, earlier work has focused on only a few transformations, such as image translation, rotation, or coloring. We close this gap by analyzing and improving robustness against twenty-four different physical transformations. First, we demonstrate that adversarial attacks based on each physical transformation significantly reduce the accuracy of deep neural networks. Next, we achieve robustness against these attacks based on adversarial training, where we show that single-step data augmentation significantly improves robustness against these attacks. We also demonstrate the generalization of robustness to these types of attacks, where robustness achieved against one attack also generalizes to some other attack vectors. Finally, we show that using an ensemble-based robust training approach, robustness against multiple attacks can be achieved simultaneously by a single network. In particular, our proposed method improves the aggregate robustness, against twenty-four different attacks, from 21.4% to 50.0% on the ImageNet dataset.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"25 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":"116748191","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/MILCOM52596.2021.9652980
Jae Y. Park, Florendo S. Encarnado, C. R. Flint, K. Redwine
Recent expansion of wireless radio frequency (RF) communications services to the maritime environment has attracted increased attention to propagation and channel modeling over the sea surface. The study of RF maritime propagation has a long history, but much research is still needed to characterize and model the ship-to-ship communication channel. In this paper, the phenomenon of the ship-to-ship channel being frequency selective is introduced. It is demonstrated that this frequency selectivity depends on signal bandwidth and link range, but not on delay spread, a metric typically applied in classifying terrestrial fading channels. Another distinctive feature of this channel is that the sea surface-reflected multipath delay is less than 1 nanosecond. Taking the extremely short delay into account, modeling of the ship-to-ship channel is provided for waveform performance analysis. Performance of coherent binary phase-shift keying (BPSK) is evaluated using the model developed. Results show that the BPSK performance degradation due to maritime fading can be substantial, especially for wideband signals, producing irreducible errors.
{"title":"Characterization and Modeling of Frequency-Selective Ship-to-Ship Channels","authors":"Jae Y. Park, Florendo S. Encarnado, C. R. Flint, K. Redwine","doi":"10.1109/MILCOM52596.2021.9652980","DOIUrl":"https://doi.org/10.1109/MILCOM52596.2021.9652980","url":null,"abstract":"Recent expansion of wireless radio frequency (RF) communications services to the maritime environment has attracted increased attention to propagation and channel modeling over the sea surface. The study of RF maritime propagation has a long history, but much research is still needed to characterize and model the ship-to-ship communication channel. In this paper, the phenomenon of the ship-to-ship channel being frequency selective is introduced. It is demonstrated that this frequency selectivity depends on signal bandwidth and link range, but not on delay spread, a metric typically applied in classifying terrestrial fading channels. Another distinctive feature of this channel is that the sea surface-reflected multipath delay is less than 1 nanosecond. Taking the extremely short delay into account, modeling of the ship-to-ship channel is provided for waveform performance analysis. Performance of coherent binary phase-shift keying (BPSK) is evaluated using the model developed. Results show that the BPSK performance degradation due to maritime fading can be substantial, especially for wideband signals, producing irreducible errors.","PeriodicalId":187645,"journal":{"name":"MILCOM 2021 - 2021 IEEE Military Communications Conference (MILCOM)","volume":"41 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":"117214038","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：