IoT for Defense and National Security

Cover
Title Page
Copyright
Contents
List of Contributors
Introduction: IoT for Defense and National Security
Part 1 Introduction: Vision, Applications, and Opportunities
	Chapter 1 Internet of Battlefield Things: Challenges, Opportunities, and Emerging Directions
		1.1 IoBT Vision
		1.2 IoBT vs. IoT
		1.3 IoBT Operational Requirements
		1.4 An Organizing Concept
			1.4.1 The MDO Effect Loop
			1.4.2 Technical Challenges
				1.4.2.1 Compositionality and Synthesis
				1.4.2.2 Timeliness and Efficiency
				1.4.2.3 Robustness to Adversarial Disruption
				1.4.2.4 Deployability at the Point of Need
		1.5 Performant and Resilient IoBTs
			1.5.1 Compositionality and Synthesis
			1.5.2 Timeliness and Efficiency
			1.5.3 Robustness to Adversarial Disruption
			1.5.4 Deployability at the Point of Need
		1.6 Future Directions
			1.6.1 Multi‐tenancy and Multiplicity of Use
			1.6.2 Multiplicity of Function
			1.6.3 Non‐stationarity and Multiplicity of Perturbations
			1.6.4 Multiplicity of Sensing Modalities
			1.6.5 Multiplicity of Time‐scales
			1.6.6 Architecture
		1.7 Conclusion
		References
	Chapter 2 Sensorized Warfighter Weapon Platforms: IoT Making the Fog of War Obsolete
		2.1 Introduction
		2.2 IoT for Firearms
		2.3 New Insights into the Battlefield Provided by IoT
		2.4 Challenges for IoT in Soldier Weapons
		2.5 Battlefield Challenges to Aggregating and Exfiltrating Data
		2.6 Protection and Security for IoT Data Communication
		2.7 State of the Art
		2.8 Conclusion
		References
	Chapter 3 IoBT Resource Allocation via Mixed Discrete and Continuous Optimization
		3.1 Introduction
		3.2 Lattices and Submodular Functions
		3.3 Problem Formulation
		3.4 An Equivalent Parameterization
		3.5 Returning to Constraints
			3.5.1 Knapsack Constraints
			3.5.2 Continuous Budget Constraints
		3.6 Computational Examples
			3.6.1 Unconstrained Optimization
			3.6.2 Knapsack‐Constrained Allocations
			3.6.3 Continuous Budget‐constrained Allocations
		3.7 Conclusions
		References
	Chapter 4 Operationalizing IoT Data for Defense and National Security
		4.1 Introduction
		4.2 Problem Statement
		4.3 Challenges
		4.4 Security Considerations
		4.5 Developing a Strategy for Operationalizing Data
		4.6 Precedence
		4.7 End State
		4.8 Conclusion
		References
	Chapter 5 Real Time Monitoring of Industrial Machines using AWS IoT
		5.1 Problem Statement
		5.2 Solution Statement – Overview
		5.3 Solution Statement – Edge Computing
		5.4 Solution Statement – Cloud Connectivity
		5.5 Solution Statement – Streaming Analytics and Data Storage
		5.6 Solution Statement – Data Visualization
		5.7 Solution Statement – Example Data Visualizations
		5.8 Results
		5.9 Next Steps
		References
	Chapter 6 Challenges and Opportunities of IoT for Defense and National Security Logistics
		6.1 Introduction
		6.2 Linking Industry and DoD Uses of IoT
		6.3 Situational Awareness
			6.3.1 Policy and Legal Implications
			6.3.2 Challenges and Considerations
		6.4 Applications for DoD
			6.4.1 Situational Awareness of People and Equipment for Maintainability and Sustainability
			6.4.2 Data Collection for Real‐time and Predictive CBM
			6.4.3 Prepositioning and Planning for People and Supplies (Prepo‐in‐motion)
			6.4.4 IoT at DoD Installations
				6.4.4.1 Energy Management
				6.4.4.2 Installations as Training Platforms
			6.4.5 IoT and Emergency Response
			6.4.6 IoT and Disaster Response
		6.5 Observations on the Future
		Acknowledgement
		References
	Chapter 7 Digital Twins for Warship Systems: Technologies, Applications and Challenges
		7.1 Introduction
		7.2 A Digital Twin Architecture for Implementation
			7.2.1 Physical Level
			7.2.2 Physical World/Virtual World Interface
			7.2.3 Digital Twin
				7.2.3.1 Integration of Functionalities: User Interfaces
				7.2.3.2 Simulation Models
				7.2.3.3 Data Storage and Data Lakes
				7.2.3.4 Data Analysis, Machine Learning, and Predictive Algorithms
		7.3 Ship Digital Twin Implementation
			7.3.1 Physical Level
			7.3.2 Physical World/Virtual World Interface
			7.3.3 Integration of Functionalities and the User Interface
			7.3.4 Simulation Models
			7.3.5 Data Analysis, Machine Learning, and Predictive Algorithms
		References
Part 2 Introduction: Artificial Intelligence and IoT for Defense and National Security
	Chapter 8 Principles of Robust Learning and Inference for IoBTs
		8.1 Internet of Battlefield Things and Intelligence
		8.2 Dimensions of Responsible AI
			8.2.1 Research Challenges in IoBTs
			8.2.2 Trust, Resilience and Interpretability
		8.3 Detecting Surprise: Adversarial Defense and Outlier Detection
		8.4 Novel Deep Learning Representation: Dynamical System
		8.5 Robust Secure State Estimation
		8.6 Distributionally Robust Learning
		8.7 Future Directions
		8.8 Conclusion
		References
	Chapter 9 AI at the Edge: Challenges, Applications, and Directions
		9.1 Introduction
		9.2 IoT Applications
			9.2.1 Visual Inspection of Assets
				9.2.1.1 Visual Recognition
				9.2.1.2 AI Optimization
				9.2.1.3 Fixed IoT Sensors vs. RIDs
			9.2.2 Thermal Inspection of Assets
				9.2.2.1 Inspection at Electric Substations
				9.2.2.2 Proposed Automation
			9.2.3 Inspection of Analog Meters and Gauges
				9.2.3.1 Gauge Detection
				9.2.3.2 Perspective Correction
				9.2.3.3 Pointer Detection and Text Recognition
			9.2.4 Other Defense and Commercial Use Cases
		9.3 Distributed AI Architecture
			9.3.1 Background: Centralized AI and Edge AI
				9.3.1.1 Centralized AI
				9.3.1.2 Edge AI
			9.3.2 Open Challenges in Edge AI
			9.3.3 New Paradigm: Distributed AI
		9.4 Technology
			9.4.1 Data Ops
				9.4.1.1 Statistical Summaries
				9.4.1.2 Dimensionality Reduction
				9.4.1.3 Sampling from Original Space
			9.4.2 Model Ops
				9.4.2.1 OOD Detection Algorithm
				9.4.2.2 Experiments
			9.4.3 Optimization and Adaptation
				9.4.3.1 Model Pruning
				9.4.3.2 Model Quantization
				9.4.3.3 Other Schemes
				9.4.3.4 Experiments: Model Optimization for Asset Inspection
			9.4.4 Federated Learning
				9.4.4.1 Resource Efficiency of FL
				9.4.4.2 Privacy Considerations
		9.5 Research Directions
			9.5.1 Learning with Resource Optimization
			9.5.2 Collaboration Among Humans and Robots
			9.5.3 Multi‐modal Learning
				9.5.3.1 Context‐based Multi‐modal Sensing
				9.5.3.2 Adaptive Navigation to Optimize Sensing
		9.6 Conclusions
		References
	Chapter 10 AI Enabled Processing of Environmental Sounds in Commercial and Defense Environments
		10.1 Introduction
			10.1.1 Challenges
			10.1.2 System Overview
			10.1.3 IoT Acoustics vs. Speech Recognition
		10.2 Use Cases
			10.2.1 Defense Use Cases
				10.2.1.1 Perimeter Defense
				10.2.1.2 Vehicle Classification
				10.2.1.3 Activation of Other Modalities
				10.2.1.4 Fleet and Facilities Maintenance
			10.2.2 Commercial Use Cases
				10.2.2.1 Manufacturing
				10.2.2.2 Vehicle Monitoring
				10.2.2.3 Animal Husbandry
				10.2.2.4 Healthcare
				10.2.2.5 Security
		10.3 System Architecture
		10.4 Technology
			10.4.1 Data Management and Curation
			10.4.2 Model Training Pipeline
			10.4.3 Models
				10.4.3.1 Shallow Models
				10.4.3.2 Deep Models
				10.4.3.3 Inference Performance on the Edge
			10.4.4 Anomaly Detection
			10.4.5 Model Drift
			10.4.6 Model Update/Evolution
			10.4.7 Model Adaptation
		10.5 Summary
		References
Part 3 Introduction: Security, Resiliency, and Technology for Adversarial Environments
	Chapter 11 Assurance by Design for Cyber‐physical Data‐driven Systems
		11.1 Introduction
			11.1.1 Formal Methods for Software Intensive Systems
			11.1.2 Adapting Formal Methods for Data Driven Systems
		11.2 Methods for Assurance
			11.2.1 Tools for Information Freshness
			11.2.2 Methods for Decision Assurance
				11.2.2.1 Scenario Generation for CPDDSs
				11.2.2.2 Consequence Assessment for CPDDSs
			11.2.3 Assurance of Interconnected Networked CPDDSs
				11.2.3.1 Network Representation
				11.2.3.2 Dynamic Cascade Modeling
				11.2.3.3 Multi‐Agent Decision Optimization
		11.3 Discussion and Conclusion
		References
	Chapter 12 Vulnerabilities in IoT Systems
		12.1 Introduction
			12.1.1 IoT System Components
			12.1.2 Vulnerabilities and Threats
				12.1.2.1 Devices
				12.1.2.2 Communication Protocols
				12.1.2.3 IoT Applications
				12.1.2.4 Physical Medium
				12.1.2.5 Mobile Apps
		12.2 Firmware
			12.2.1 Unprotected Network Services
			12.2.2 Unprotected Firmware Updating
			12.2.3 Buffer Overflow
		12.3 Communication Protocols
			12.3.1 Wi‐Fi
			12.3.2 Zigbee
			12.3.3 Z‐Wave
			12.3.4 Bluetooth
			12.3.5 Physical Layer
				12.3.5.1 Jamming Attack
				12.3.5.2 Side Channel Attack
			12.3.6 TCP/IP Suite & Application Layer
		12.4 IoT Apps
			12.4.1 Checking Safety and Security Properties
			12.4.2 Dynamic Security Policy Enforcement
			12.4.3 IoT App Sniffing
		12.5 Physical Dependencies
		12.6 Companion Mobile Apps
		12.7 Hardware
		12.8 IoT Platforms
			12.8.1 Over‐privileging
			12.8.2 Data Leakage
		12.9 Countermeasures
		12.10 Conclusions
		References
	Chapter 13 Intrusion Detection Systems for IoT
		13.1 Introduction
		13.2 Background
			13.2.1 Intrusion Detection Systems
				13.2.1.1 Placement of Collectors
				13.2.1.2 Architecture of Analyzers
				13.2.1.3 Detection Mechanisms
			13.2.2 Characteristics of IoT Environments
				13.2.2.1 Simple Networking Patterns
				13.2.2.2 Diverse Network Protocols
				13.2.2.3 Small Number of Threads
				13.2.2.4 Various Types of CPU Architectures and Operating Systems
				13.2.2.5 Resource Constraints
				13.2.2.6 Large Numbers of Devices
				13.2.2.7 Dynamics and Autonomy
			13.2.3 IoT‐Specific Protocols
				13.2.3.1 IoT Network‐layer Protocols
				13.2.3.2 IoT Application‐layer Protocols
			13.2.4 IDS in IoT Environment
				13.2.4.1 Relevance of IDS in IoT Environment
				13.2.4.2 Challenges for IDSes in IoT Dynamic and Autonomous Environment
		13.3 IoT Attack Scenarios
			13.3.1 Attacks from the Internet
				13.3.1.1 Port Scanning
				13.3.1.2 Telnet/SSH/HTTP Bruteforce
				13.3.1.3 SYN/ACK/UDP/HTTP Flooding
			13.3.2 IoT‐specific Network‐layer Attacks
				13.3.2.1 Hello Flood Attack
				13.3.2.2 Neighbor Attack
				13.3.2.3 DIS Attack
				13.3.2.4 Sinkhole Attack
				13.3.2.5 Wormhole Attack
				13.3.2.6 Grayhole (or Selective Forwarding) Attack
			13.3.3 IoT‐specific Application‐layer Attacks
				13.3.3.1 CONNECT/CONNACK Flooding
				13.3.3.2 CoAP Request/ACK Flooding
		13.4 Proposed IDSes for IoT
			13.4.1 Definition of Normal/Abnormal Behavior
				13.4.1.1 Legitimate IP Addresses
				13.4.1.2 Threshold
				13.4.1.3 Automata
				13.4.1.4 Federated Learning
			13.4.2 Enhancements of ML‐based Detectors
				13.4.2.1 Compression Header Analyzer Intrusion Detection System (CHA‐IDS)
				13.4.2.2 E‐Spion
				13.4.2.3 Deep learning-based IDS (DL‐IDS)
				13.4.2.4 Multiclass Classification Procedure
				13.4.2.5 Discussion
			13.4.3 Lightweight Detector Implementation
				13.4.3.1 Raspberry Pi IDS (RPiDS)
				13.4.3.2 Passban IDS
				13.4.3.3 Discussion
			13.4.4 Combination of Diverse Detectors
				13.4.4.1 IDS with Game‐theoretic Methodology
				13.4.4.2 Hybrid Intrusion Detection and Prevention System (IDPS)
				13.4.4.3 IDPS
				13.4.4.4 Discussion
			13.4.5 Optimal Detector Selection
				13.4.5.1 Kalis
				13.4.5.2 Reinforcement learning-based IDS (RL‐IDS)
				13.4.5.3 Discussion
		13.5 Research Directions
		Acknowledgement
		References
	Chapter 14 Bringing Intelligence at the Network Data Plane for Internet of Things Security
		14.1 Introduction
		14.2 Related Work
		14.3 System Design
			14.3.1 Architecture of the FRG Approach
			14.3.2 Architecture of the BNN Approach
		14.4 Problem Modeling
			14.4.1 Classification with Header Bytes
			14.4.2 Classification with Header Fields
		14.5 Algorithms and Learning Models
			14.5.1 FRG Approach: Overview
			14.5.2 FRG Stage 1: Neural Network Structure
			14.5.3 FRG Stage 2: Header Field Definition
			14.5.4 BNN Approach
		14.6 Evaluation Results
			14.6.1 Performance of FRG Approach: Setup and Metrics
			14.6.2 Performance of FRG Stage 1 (Classification)
			14.6.3 Performance of FRG Stage 2 (Header Field Definition)
				14.6.3.1 Profiles of Importance Scores
				14.6.3.2 Impact of Header Fields on Accuracy
				14.6.3.3 Impact of Header Fields on Costs
				14.6.3.4 Optimal Selection of Header Fields
			14.6.4 Performance of BNN Approach
				14.6.4.1 Main Takeaways
		14.7 Conclusions and Future Challenges
		Acknowledgment
		References
	Chapter 15 Distributed Computing for Internet of Things Under Adversarial Environments
		15.1 Introduction
		15.2 Distributed Computing for IoT in Defense Applications
			15.2.1 Overview of Requirements/Challenges
			15.2.2 Characteristics of Distributed IoBT Applications
		15.3 Threat Model
			15.3.1 System Description
			15.3.2 Threats
				15.3.2.1 Goals of an Adversary
				15.3.2.2 Attack Vectors
		15.4 Frameworks for Distributed Computing
			15.4.1 Resource and Task Management in Distributed Computing
			15.4.2 Gathering Resources in Adversarial Environments
		15.5 Establishing Trust in Adversarial Environments: Solutions and Open Opportunities
			15.5.1 Verifiable Computation
				15.5.1.1 Homomorphic Encryption
				15.5.1.2 Proof‐based Verification
				15.5.1.3 TrueBit
				15.5.1.4 Perlin
				15.5.1.5 Open Opportunities
			15.5.2 Byzantine Fault‐tolerant Distributed Computing
				15.5.2.1 Open Opportunities
			15.5.3 Grey Resource Accumulation
				15.5.3.1 Open Opportunities
			15.5.4 Cryptographic Approaches
				15.5.4.1 Open Opportunities
			15.5.5 Secure Computation with Trusted Execution Environments
				15.5.5.1 Open Opportunities
		15.6 Summary
		Acknowledgment
		References
	Chapter 16 Ensuring the Security of Defense IoT Through Automatic Code Generation
		16.1 The Challenge of IoT in Defense and National Security Applications: The Challenge
		16.2 Solutions
			16.2.1 Control the Interfaces Between IoT Elements
			16.2.2 Problems with Traditional Approaches to Malware Protection
			16.2.3 Traditional Approaches to Security: Hardware
			16.2.4 Traditional Approaches to Security: Simulation
			16.2.5 Traditional Approaches to Security: Software
				16.2.5.1 Coding Weaknesses, Software Vulnerabilities and Malware
				16.2.5.2 Traditional Approaches for Protecting IoT Software
				16.2.5.3 Improvements on Traditional Software Approaches
			16.2.6 Auto‐code Generation for Vulnerability‐free IoT
				16.2.6.1 Applying Auto‐code Generation Selectively for IoT Network Security
				16.2.6.2 A Practical Approach to Generating Vulnerability‐free IoT Networks
		16.3 Automatic Code Generation
			16.3.1 Core Auto‐generation Engine
			16.3.2 Semantic Definitions of Software Functions
			16.3.3 Formal Methods for Verifying Semantic Definitions
				16.3.3.1 Static Analysis for Verifying Code Generator Produces Vulnerability‐free Code
			16.3.4 An Extended Example: Automatic Generation of Router Software
		16.4 IoT Interface‐code Issuing Authority
			16.4.1 Role of IoT Interface‐code Authority (IICA)
			16.4.2 Precedents and Examples and a Proposed IoT Interface Code Authority
		16.5 Conclusions
		References
Part 4 Introduction: Communications and Networking
	Chapter 17 Leveraging Commercial Communications for Defense IoT
		17.1 Introduction
		17.2 Key Differences Between Defense and Commercial Communications Requirements
			17.2.1 Interoperability
			17.2.2 Mobility
			17.2.3 Security
			17.2.4 Vulnerability
		17.3 Key Differences Between Defense and Commercial Technology Development
		17.4 Commercial Communications for Use in Defense and Homeland Security
		17.5 Conclusion
		References
	Chapter 18 Military IoT: Tactical Edge Clouds for Content Sharing Across Heterogeneous Networks
		18.1 Introduction
		18.2 The Need for Tactical Edge Clouds
		18.3 Two Architectures
			18.3.1 Architecture Paradigm 1: DARPA CBMEN
			18.3.2 Architecture Paradigm 2: DARPA DyNAMO
		18.4 Tactical Edge Cloud Architectural Insights
			18.4.1 Information Generation and Discovery
			18.4.2 Information Availability
			18.4.3 Controlling Access
			18.4.4 Information Quality of Service
			18.4.5 Information Importance
		18.5 Summary
		Acknowledgment
		References
	Chapter 19 Spectrum Challenges in the Internet of Things: State of the Art and Next Steps
		19.1 Introduction
		19.2 Spectrum Bands of Interest in the Internet of Things
			19.2.1 Low‐bands and Mid‐bands
				19.2.1.1 Millimeter‐Wave Bands
				19.2.1.2 Visible Light and Communications Above 100 GHz
		19.3 Spectrum Management in the Internet of Things: Requirements and Existing Work
		19.4 Spectrum Management in the Internet of Things: The Way Ahead
			19.4.1 Protecting Passive and Incumbent Users from IoT Interference in Shared Bands
			19.4.2 Experimental Spectrum Sharing at Scale Through the Colosseum and NSF PAWR Testbeds
			19.4.3 Robust Machine Learning for Effective, Reliable and Efficient Spectrum Management
			19.4.4 The Role of O‐RAN in Spectrum Sharing
		19.5 Conclusions
		References
	Chapter 20 Tactical Edge IoT in Defense and National Security
		20.1 Introduction
		20.2 Background
			20.2.1 Tactical Edge IoT drivers
			20.2.2 Defense and Public Safety
		20.3 Compelling COTS Edge IoT Applications
		20.4 Target Scenarios for Tactical Edge IoT
			20.4.1 C4ISR
			20.4.2 Firepower Control Systems
			20.4.3 Logistics
				20.4.3.1 Fleet Management
				20.4.3.2 Individual Supplies
			20.4.4 Smart City Operations
			20.4.5 Soldier Healthcare and Workforce Training
			20.4.6 Collaborative and Crowd Sensing
			20.4.7 Energy Management
			20.4.8 Smart Surveillance
		20.5 Communications Architecture
		20.6 Main Challenges and Recommendations
		20.7 Conclusions
		Acknowledgments
		References
	Chapter 21 Use and Abuse of IoT: Challenges and Recommendations
		21.1 The Elements of IoT and Their Nature
			21.1.1 Use and Abuse of IoT
				21.1.1.1 What Makes IoT So Powerful?
				21.1.1.2 Orwell's Vision Has Not Yet Fully Materialized
				21.1.1.3 IoT Unites Sensing/Information‐Extraction with Intelligent Processing and Action
			21.1.2 Pervasive Sensing and Information Extraction
				21.1.2.1 Sensors and Sensor Networks
				21.1.2.2 Information Extraction
			21.1.3 Intelligent Processing
				21.1.3.1 IoT and the Nature of Intelligent Processing (AI)
				21.1.3.2 Intelligent Processing of IoT Sensor Data and Extracted Information
				21.1.3.3 Abuses of IoT Arising from Problems with Intelligent Processing
			21.1.4 Control of Actions by IoT Devices
				21.1.4.1 Control of Action
				21.1.4.2 Abuse of Action by IoT
		21.2 Preventing the Abuse of IoT While Enabling Its Benefits
			21.2.1 A General Framework
				21.2.1.1 The Need and Basis for an IoT Framework to Protect Human Rights
				21.2.1.2 Consent by the Public and the Governed
				21.2.1.3 Transparency: The Foundation of Consent
				21.2.1.4 Accountability and Consequences
				21.2.1.5 Security and Integrity
		21.3 Types of Abuse and Misuse, and Prevention Through Regulation
			21.3.1 Types of Abuse of IoT
				21.3.1.1 Type 1 Abuse: Illegal or Unethical Abuse by Individuals or Organizations
				21.3.1.2 Type 2 Abuse: Legal Abuse of IoT Without Consent or Benefit to Users or Owners
				21.3.1.3 Type 3 Abuse: Government Abuse While Using IoT for Public Defense, Health, Safety, and Wellbeing
				21.3.1.4 Type 4 Abuse: Government Use of IoT to Enhance Its Own Power and Enrich Officials
			21.3.2 Regulating IoT to Prevent Abuse While Advancing Its Benefits
				21.3.2.1 The Right to Limit and Regulate IoT
				21.3.2.2 Regulating IoT: A Summary
		21.4 Concluding Remarks: A Call to Action
		References
Index
EULA