Handbook of Dynamic Data Driven Applications Systems: Volume 2

✍ Scribed by Frederica Darema (editor), Erik P. Blasch (editor), Sai Ravela (editor), Alex J. Aved (editor)

• Publisher: Springer
• Year: 2023
• Tongue: English
• Leaves: 937
• Category: Library

✦ Synopsis

This Second Volume in the series Handbook of Dynamic Data Driven Applications Systems (DDDAS) expands the scope of the methods and the application areas presented in the first Volume and aims to provide additional and extended content of the increasing set of science and engineering advances for new capabilities enabled through DDDAS. The methods and examples of breakthroughs presented in the book series capture the DDDAS paradigm and its scientific and technological impact and benefits. The DDDAS paradigm and the ensuing DDDAS-based frameworks for systems’ analysis and design have been shown to engender new and advanced capabilities for understanding, analysis, and management of engineered, natural, and societal systems (“applications systems”), and for the commensurate wide set of scientific and engineering fields and applications, as well as foundational areas. The DDDAS book series aims to be a reference source of many of the important research and development efforts conducted under the rubric of DDDAS, and to also inspire the broader communities of researchers and developers about the potential in their respective areas of interest, of the application and the exploitation of the DDDAS paradigm and the ensuing frameworks, through the examples and case studies presented, either within their own field or other fields of study.

As in the first volume, the chapters in this book reflect research work conducted over the years starting in the 1990’s to the present. Here, the theory and application content are considered for:

• Foundational Methods
  
• Materials Systems
  
• Structural Systems
  
• Energy Systems
  
• Environmental Systems: Domain Assessment & Adverse Conditions/Wildfires
  
• Surveillance Systems
  
• Space Awareness Systems
  
• Healthcare Systems
  
• Decision Support Systems
  
• Cyber Security Systems
  
• Design of Computer Systems
  

The readers of this book series will benefit from DDDAS theory advances such as object estimation, information fusion, and sensor management. The increased interest in Artificial Intelligence (AI), Machine Learning and Neural Networks (NN) provides opportunities for DDDAS-based methods to show the key role DDDAS plays in enabling AI capabilities; address challenges that ML-alone does not, and also show how ML in combination with DDDAS-based methods can deliver the advanced capabilities sought; likewise, infusion of DDDAS-like approaches in NN-methods strengthens such methods. Moreover, the “DDDAS-based Digital Twin” or “Dynamic Digital Twin”, goes beyond the traditional DT notion where the model and the physical system are viewed side-by-side in a static way, to a paradigm where the model dynamically interacts with the physical system through its instrumentation, (per the DDDAS feed-back control loop between model and instrumentation).

✦ Table of Contents

Contents
About the Editors
1 The Dynamic Data Driven Applications Systems (DDDAS) Paradigm and Emerging Directions
1.1 Introduction
1.2 Overview of the DDDAS Paradigm
1.2.1 The DDDAS Paradigm: Definition and Features
1.2.2 DDDAS Background
1.2.3 State Estimation and Data Assimilation
1.2.4 DDDAS and Adaptive State Estimation
1.2.5 Does DDDAS Use Feedback Control?
1.3 DDDAS Methods
1.4 Historical Perspective of DDDAS Research and Breakthroughs
1.4.1 Theory: Modeling and Analysis
1.4.2 Methods: Domain Applications
1.4.3 Analysis and Design: Systems and Architectures
1.5 Emerging Domains of DDDAS Impact
1.5.1 DDDAS and Data Assimilation and Digital Twins
1.5.2 DDDAS and Test and Evaluation
1.5.3 DDDAS and 5G and Beyond Networks
1.6 Differences and Advantages of DDDAS Vis-a-Vis “AI” Methods
1.6.1 DDDAS and AI Techniques
1.6.2 DDDAS Inspires NN Methods
1.6.3 Differences and Advantages of DDDAS Versus CPS
1.7 Book Overview
1.8 Summary
References
Part I Fundamentals Aware: Theory/Foundational Methods
2 Dynamic Data-Driven Applications Systems and Information-Inference Couplings
2.1 Introduction
2.2 System Dynamics and Optimization
2.2.1 DDDAS and Sparsity: An Example
2.3 DDDAS: From Inference to Information and Back
2.3.1 Human-Machine Teaming: An Example
2.4 Emerging Opportunities
2.5 Conclusions
References
3 Polynomial Chaos Expansion-Based Nonlinear Filtering for Dynamic State Estimation
3.1 Introduction
3.2 Literature Review
3.3 Uncertainty Propagation by Using Polynomial Chaos Expansion
3.4 Polynomial Chaos-Based Ensemble Square Root Filter
3.4.1 Ensemble Square Root Filter
3.4.2 Polynomial Chaos-Based Ensemble Square Root Filter
3.5 Existing Nonlinear Filters and Complexity Analysis
3.6 Application to Ballistic Trajectory Estimation Problems
3.7 Conclusion
References
4 Measure-Invariant Symbolic Systems for Pattern Recognition and Anomaly Detection
4.1 Introduction
4.2 Background
4.3 Symbolic Time Series Analysis
4.3.1 Measure-Invariant Symbolic Systems
4.3.2 Probabilistic Finite State Automata
4.3.3 D-Markov Machines
4.3.4 Anomaly Detection in the STSA Setting
4.4 Technical Approach
4.5 Experimental Validation: Results and Discussion
4.5.1 Application #1: TAI Emulation in Combustion Systems
4.5.2 Application #2: Fatigue Damage in a Polycrystalline Alloy
4.5.3 Comparison of CPU Execution Time for Both Applications
4.5.4 Threshold Parameters τ1 and τ2 in Algorithm 1
4.6 Summary, Conclusions, and Future Research
References
Part II Materials Systems
5 Equation-Free Computations as DDDAS Protocols for Bifurcation Studies: A Granular Chain Example
5.1 Introduction
5.1.1 Overview
5.1.2 Dark Breathers in Engineered Granular Chains
5.1.3 Equation-Free Bifurcation Studies
5.2 Theoretical Method and the Experimental Setup
5.2.1 The EF/DDDAS Approach
5.2.1.1 Numerically Computing Solution Branches
5.2.1.2 Linear Stability Analysis
5.2.1.3 Estimating the Unstable Manifold
5.2.2 Experimental Setup
5.3 An Equation-Aware Bifurcation Study
5.3.1 Governing Equations for a Damped-Driven Granular Chain
5.3.2 Model Validation
5.4 An EF/DDDAS Study of the Dynamics
5.4.1 Pitchfork Bifurcations
5.4.2 Period-Doubling Bifurcations
5.4.3 Quasi-periodic Solutions
5.4.4 Computing Unstable Manifolds
5.5 Conclusion
References
6 A Stochastic Dynamic Data-Driven Framework for Real-Time Prediction of Materials Damage in Composites
6.1 Introduction
6.2 Continuum Damage Mechanics
6.3 Experimental Measurements
6.4 A Bayesian Framework for Predictive Physics-Based Modeling
6.4.1 Model Validation and Prediction Under Uncertainty
6.5 A Stochastic Dynamic Data-Driven Applications Systems
6.6 Results
6.6.1 Statistical Calibration and Plausibility of Damage Models
6.6.2 State Monitoring with Bayesian Filter
6.7 Conclusions
References
7 Dynamic Data-Driven Monitoring of Nanoparticle Self-Assembly Processes
7.1 Introduction
7.2 Background
7.2.1 Nanoparticles and Nanoparticle Self-Assembly Process
7.2.2 Online of Nanometrology
7.3 Modeling Nanoparticle Self-Assembly Processes
7.3.1 Shape Modeling
7.3.2 Modeling a Nanoparticle Self-Assembly Process
7.3.3 Online Model Update with TEM Data
7.4 Adaptive TEM Triggering
7.4.1 General DLS Theory and Theoretical DLS Calculation
7.4.2 TEM Triggering
7.5 Application to Gold Nanorod Experiments
7.5.1 Initialization of the Model
7.5.2 Running the Adaptive TEM Triggering
7.5.3 Comparison to Benchmark Models
7.6 Discussion
References
Part III Structural Systems - Structural – Infrastructures
8 From Data to Decisions: A Real-Time Measurement–Inversion–Prediction–Steering Framework for Hazardous Events and Health Monitoring
8.1 Introduction
8.2 Methodology
8.2.1 Bayesian Inversion
8.2.2 Uncertainty Propagation
8.2.3 Sensor Steering
8.2.4 Model Reduction
8.3 Application of the MIPS Framework: Airborne Contaminant Dispersion
8.3.1 Overview
8.3.2 Contaminant Dispersion Simulation
8.3.3 Measurement–Inversion–Prediction–Steering for Airborne Contaminant Dispersion
8.4 Airborne Contaminant Dispersion: Results
8.4.1 Problem Description
8.4.2 Forward Problem
8.4.3 Inverse Problem
8.4.4 Sensor Steering
8.4.5 Moving Window Illustration
8.5 Application of the MIPS Framework: Self-Aware Aerospace Vehicle
8.5.1 Self-Aware Aerospace Vehicles
8.5.2 Offline Libraries for Self-Aware Aerospace Vehicles
8.5.3 Measurement-Inversion-Prediction-Steering for Self-Aware Aerospace Vehicles
8.6 Self-Aware Aerospace Vehicle: Results
8.6.1 Measurement, Inversion, and Prediction
8.6.2 Steering
8.6.3 Demonstration
8.7 Conclusions
References
9 Bayesian Computational Sensor Networks: Small-Scale Structural Health Monitoring
9.1 General Overview
9.2 State-of-the-Art, Challenges, and Method
9.3 Robot Monitoring Agents
9.3.1 The SLAMBOT
9.3.2 SLAM
9.4 Ultrasound Range Sensor
9.5 Data Routing Model for Distributed Cloud Computing
9.6 Validation Experiments
9.7 Conclusions and Future Work
9.8 Summary
References
10 Dynamic Data Driven Sensor Tasking with Applications in Space and Aerospace Systems
10.1 Introduction
10.1.1 DDDAS Paradigm
10.1.2 State of the Art
10.2 Problem Statement
10.3 Measures to Quantify Sensor Performance
10.4 Sensor Tasking
10.4.1 Sequential in Time
10.4.2 Sequential in Sensors
10.4.3 Sequential in Objects
10.4.4 Receding Horizon
10.4.5 Greedy Search
10.5 Optimal Sensor-Tasking Examples
10.5.1 Example of Tracking UAVs
10.5.2 Space Situational Awareness Example
10.5.2.1 Case 1: 15 Satellites and 2 Sensors
10.5.2.2 Case 2: Tracking 100 Satellites with 3 Sensors, Greedy in Targets
10.5.3 Maritime Target Tracking Example
10.6 Conclusion
References
Part IV Energy Systems – Energy Production and Distribution
11 Dynamic Data-Driven Application Systems for Reservoir Simulation-Based Optimization: Lessons Learned and Future Trends
11.1 Introduction
11.2 Optimal Reservoir Management
11.2.1 Optimization Challenges
11.2.2 Closed-Loop Reservoir Management
11.2.3 System Architecture
11.3 Summary of Previous Work
11.3.1 Reservoir Simulation: IPARS
11.3.1.1 Multiblock Simulations
11.3.1.2 Computational Support for Multiblock Coupling
11.3.1.3 Optimization Methods
11.3.2 Computational Collaboratory: DISCOVER
11.3.2.1 Autonomic Grid Middleware for Coupling of Simulations, Data, and Resources
11.3.2.2 The Discover Computational Collaboratory (DCC)
11.3.3 Data Exploration and Analysis: DataCutter
11.3.3.1 Data Exploration Using DataCutter
11.3.3.2 Experimental Performance Evaluation
11.4 Extensions to Previous Work
11.4.1 Extensions to the Physics Simulator IPARS
11.4.1.1 New Physics Capabilities
11.4.1.2 New Discretization and Meshing Schemes
11.4.1.3 New Optimization and History-Matching Capabilities
11.4.2 Extensions to Multiblock Coupling Infrastructure: DataSpaces
11.4.3 Extensions to Computational Collaboratory
11.4.4 Extensions to DataCutter
11.5 Key Findings and Lessons Learned
11.5.1 On Reservoir Simulations and Coupled Physics Modeling
11.5.1.1 Mass Conservation and Implicit Schemes for Compositional Modeling
11.5.1.2 Iterative Methods for Coupling Flow and Geomechanics
11.5.1.3 Computational Efficiency: Solvers, Adaptive Meshing, and Domain Decomposition
11.5.1.4 A Posteriori Error Estimates
11.5.1.5 Parallel History Matching and Optimization
11.5.2 On Discover and Grid-Enabled Middleware Services
11.5.3 On DataCutter and Management of Multidimensional Datasets
11.6 Current and Future Directions of DDDAS in Reservoir Simulation-Based Optimization
11.6.1 Reservoir Simulations and Modeling
11.6.1.1 Reparameterization and Dimensionality Reduction
11.6.1.2 Model Reduction
11.6.1.3 Dynamic Data-Driven Reservoir Models
11.6.2 Data Assimilation and Optimization
11.6.3 New Trends in Subsetting and Processing Datasets
11.7 Conclusions
References
12 DDDAS Within the Oil and Gas Industry
12.1 Introduction
12.2 DDDAS Across the O&G Industry
12.2.1 Upstream
12.2.1.1 Well Drilling
12.2.1.2 Well Technology
12.2.2 Midstream
12.2.3 Downstream
12.3 Shale Revolution
12.3.1 Perspective
12.3.2 LHC: Small Modular/Distributed Units and DDDAS
12.4 Conclusion
References
13 A Simulation-Based Online Dynamic Data-Driven Framework for Large-Scale Wind-Turbine Farm SystemsOperation
13.1 Introduction
13.2 Problem Statement and Overview of Methodology
13.3 Systems Dynamic: Abstraction and Modeling for the Discrete Event System Specification Wind-Turbine Farm Simulation Model (DEVS-WFSM)
13.4 Decision Support: A Stochastic Integer Programming Model for Wind-Turbine Farm O&M
13.4.1 Wind-Turbine Farm Maintenance Scheduling Two-Stage Stochastic Programming Model
13.5 Decision Support: Online Dynamic Data-Driven Framework (ODDF)
13.6 A Simulation-Based Online Dynamic Data-Driven Framework (SODDF): Integrating DEVS-WFSM and ODDF
13.7 Evaluation
13.7.1 Experimental Setup
13.7.2 Simulation Results and Discussion
13.8 Conclusions
References
Part V Environmental Systems – Conditions Assessment
14 Toward Dynamic Data-Driven Systems for Rapid Adaptive Interdisciplinary Ocean Forecasting
14.1 Introduction
14.2 Overview of Dynamic Data-Driven System Architecture
14.3 Real-Time Interdisciplinary Modeling and Forecasting and Process to Date Toward an Ocean Science DDDAS
14.3.1 Objective Adaptive Sampling Using ESSE
14.3.2 Generalized Biological Modeling and Non-Automated Physical Adaptation
14.3.2.1 Monterey Bay Application
14.3.3 Acoustical-Biological Measurement Models
14.4 Components of the Poseidon System: Architecture Design and Progress
14.4.1 Distributed/Grid Computational Strategies
14.4.2 User Interfaces
14.4.2.1 Generic Web User Interfaces
14.4.2.2 Remote Visualization and Feature Extraction
14.5 Conclusions
References
15 Toward Cyber-Eco Systems: Networked Sensing, Inference, and Control for Ecological and Agricultural Systems
15.1 Introduction
15.2 Overview
15.2.1 Overview of Deployed Systems
15.3 Sampling, Transmission, and Inference
15.3.1 Hierarchical Bayesian Inference
15.3.2 Energy-Aware Sensing and Inference
15.4 Networking
15.4.1 Physical and Multiple Access Layers
15.4.2 Middleware
15.4.3 Cross-Layer Coding and Decoding
15.5 Node Hardware/Software Design
15.5.1 Example: Sensor/Actuator Node Design for a Cyber-eco System
15.5.2 Self-Sufficiency and Energy Neutrality
15.6 Emerging Applications
15.6.1 Closed-Loop Control
15.6.2 On-Line Optimization
15.6.3 New Applications
15.7 Conclusion
References
16 An Energy-Aware Airborne Dynamic Data-Driven Application System for Persistent Sampling and Surveillance
16.1 Introduction
16.2 The Energy-Aware Airborne Dynamic Data-Driven Application Systems
16.2.1 Dual-Doppler Synthesis
16.2.2 Atmospheric Model for Online Planning
16.2.2.1 Horizontal Wind Estimation
16.2.2.2 Vertical Velocity Estimation
16.2.2.3 Dynamic Model
16.2.3 Wind Field Database
16.2.4 Lattice Planner
16.2.5 Trajectory Optimization Layer
16.2.6 Five-Hole Probe and Sonde
16.3 Experimental Results
16.3.1 Airdata Verification and Integrated Airborne Tempest Experiment
16.3.2 Multi-sUAS Evaluation of Techniques for Measurement of Atmospheric Properties
16.3.3 Energy-Aware Airborne DDDAS (Dynamic Data-Driven Applications Systems)
16.3.3.1 Loiter Mission
16.3.3.2 Combined Mission and Stitched Trajectory
16.3.3.3 Wind Energy Harvesting
16.4 Conclusion
References
Part VI Environmental Systems – Adverse Conditions (Fire Modeling)
17 Using Dynamic Data-Driven Cyberinfrastructure for Next-Generation Wildland Fire Intelligence
17.1 Introduction
17.1.1 Modeling Fire Behavior
17.1.2 Understanding the Fire Ignition and Perimeter Growth
17.2 WIFIRE Cyberinfrastructure for Wildland Fire Science
17.2.1 Fire Modeling Workflows at the Digital Continuum
17.2.2 WIFIRE Commons
17.2.3 Firemap: A Data-Driven Decision-Support Tool for Operational Fire Monitoring and Response
17.3 Wildfire State Simulation
17.4 The Need for Data Assimilation
17.5 Linear Time-Variant Kalman Filter
17.6 Ensemble Kalman Filter
17.7 Illustration of Wildfire Data Assimilation in Firemap
17.8 Summary Remarks
References
18 Autonomous Monitoring of Wildfires with Vision-Equipped UAS and Temperature Sensors via Evidential Reasoning
18.1 Introduction
18.2 Wildfire Dynamics
18.2.1 Spatio-Temporal Spread Model: The Minimum Travel Time Method
18.3 Belief Modeling and Information Fusion
18.3.1 Fundamentals of Evidential Reasoning
18.3.2 Monte Carlo Forecaster
18.3.2.1 Construction of the Forecasting Agent's Beliefs
18.3.3 Temperature Sensor
18.3.3.1 Construction of the Temperature Agent's Beliefs
18.3.4 Vision Sensor
18.3.4.1 Construction of the Visual Agent's Beliefs
18.4 Bi-directional Feedback: Aerial Sensor Guidance Methods
18.4.1 Guidance Methods
18.4.1.1 Conflict-Guided Method
18.4.1.2 Ignorance-Guided Method
18.4.1.3 Belief-Guided Method
18.4.2 Mobile Sensor Dynamics
18.4.2.1 Drone Dynamics Mode 1: Traveling
18.4.2.2 Drone Dynamics Mode 2: Sensing and Avoidance
18.4.3 Mobile Sensor Constraints
18.4.3.1 Airspace Constraint
18.4.3.2 Battery-Life Constraint
18.4.3.3 Radiation Constraint
18.5 Complete Estimation Procedure with Bi-directional Feedback
18.6 Simulated Wildland Fire Near Taos, New Mexico
18.6.1 Large-Scale Performance Metrics
18.6.2 Simulation Setup
18.6.2.1 Simulated Ground-Truth Fires
18.6.2.2 Sensor Setup and Constraints
18.6.3 Simulation of the Estimation Scenario
18.6.4 Comparative Estimation Performance
18.6.4.1 Representation Error
18.6.4.2 Radial Error
18.7 Future Studies and Conclusion
References
19 Airborne Fire Detection and Modeling Using Unmanned Aerial Vehicles Imagery: Datasets and Approaches
19.1 Introduction
19.1.1 Problem Definition
19.1.2 Drone-Based Fire Detection and Management
19.2 Related Work
19.2.1 Related Studies
19.2.2 Related Datasets
19.3 The FLAME Dataset and Fire Classification and Modeling
19.3.1 FLAME Dataset
19.3.2 Fire Classification and Modeling Methodology
19.3.3 Fire Detection and Segmentation
19.4 Future Directions
19.5 Conclusions
References
Part VII Surveillance/Observational Systems
20 DDDAS-Based Remote Sensing
20.1 Introduction
20.2 Wildfire Modeling
20.3 Vehicle Tracking
20.3.1 Real Datasets
20.3.1.1 AeroRIT
20.3.1.2 RooftopHSI
20.3.1.3 Airborne Highway Collect
20.4 Scene Generation
20.5 Conclusions
20.6 Summary
References
21 Advances in Domain Adaptation for Aerial Imagery
21.1 Introduction
21.2 Domain Adaptation Methods
21.2.1 Adversarial Discriminative Domain Adaptation
21.2.2 SymNets
21.2.3 Open Set Domain Adaptation by Backpropagation
21.2.4 Separate to Adapt (STA)
21.3 Explainability for Domain Adaptation
21.3.1 Grad-CAM
21.4 Experiments and Results
21.4.1 Results for Unsupervised DA
21.4.2 Visualization of Results for Unsupervised DA
21.4.3 Results for Open Set DA
21.4.4 Visualization of Results for Open Set DA
21.5 Conclusion
References
Part VIII Space Awareness/Aware
22 Retrospective Cost Parameter Estimation with Application to Space Weather Modeling
22.1 Introduction
22.2 Parameter Estimation Problem
22.3 Traditional Parameter Estimation Methods
22.4 Retrospective Cost Parameter Estimation
22.4.1 Parameter Estimator
22.4.2 Retrospective Cost Optimization
22.4.3 The Filter Gf
22.5 Low-Order Example
22.6 Parameter Estimation in the Generalized Burgers Equation
22.7 Global Ionosphere-Thermosphere Model
22.8 Conclusions
References
23 A Dynamic Data-Driven Approach for Space SituationalAwareness
23.1 Introduction
23.1.1 SSA: A DDDAS-Based Approach
23.2 Joint Detection, Correlation, and Tracking
23.3 A Belief Space Perspective of RFS Bayesian Multiple Target Tracking
23.3.1 Preliminaries
23.3.2 Recursive Bayesian RFS-Based Multi-target Tracking
23.3.3 Relationship Between Classical (HOMHT) and RFS-Based Multi-target Tracking Techniques
23.3.3.1 Relationship When the Number of Targets Is Fixed
23.3.3.2 Relationship Between HOMHT and FISST with Target Birth
23.4 Smart Sampling Markov Chain Monte Carlo
23.5 Applications to Collisional Cascading (Kessler Syndrome)
23.5.1 Fragmentation with Background Objects
23.5.2 Collisional Cascading
23.6 Conclusion
References
Part IX Healthcare Systems
24 Data-Driven Cancer Research with Digital Microscopy and Pathomics
24.1 Introduction
24.2 Quantitative Digital Histopathology
24.2.1 Digital Microscopy
24.2.2 Pathomics
24.3 Quantitative Digital Histopathology in a Dynamic Data-Driven Application System Framework
24.4 Methods and Software Systems for DDDAS in Digital Pathology
24.4.1 High Performance Computing Efforts
24.4.2 Quality Control and Data Curation Methods
24.4.3 Software for Managing and Visualizing Data, Data Formats, and Database Models
24.5 Conclusions
References
25 Robust Data-Driven Region of Interest Segmentation for Breast Thermography
25.1 Introduction
25.2 Breast Thermography
25.2.1 Machine Learning for Breast Thermography
25.2.2 Imaging Protocol
25.3 Automated Breast Segmentation
25.3.1 Traditional ML Approach
25.3.1.1 Inframammary Fold Detection
25.3.1.2 Upper Boundary Detection
25.3.1.3 Body and Mid-Body Contours Detection
25.3.1.4 ROI Fitting
25.3.2 Deep Learning Approach
25.3.3 DDDAS Paradigm for Segmentation
25.4 Results
25.5 Discussion
25.6 Conclusions
References
26 Adaptive Data Stream Mining (DSM) Systems
26.1 Introduction
26.1.1 Application Examples: Healthcare and Climate Monitoring
26.1.2 Design Challenges
26.1.3 Application of DDDAS Principles
26.1.4 Broader Impacts
26.2 Design Methodologies
26.3 Application Case Study
26.3.1 Patient Healthcare System
26.3.2 Modeling and Simulation
26.3.3 Examples of DDADD System Implementations
26.3.4 Summary
26.4 Large-Scale Distributed Learning Systems
26.4.1 Mathematical Framework
26.4.2 Application Examples
26.5 Summary
References
Part X Operations Aware – Decisions Management/Optimization
27 Deception Detection in Videos Using Robust Facial Features with Attention Feedback
27.1 Introduction
27.2 Related Work
27.3 Method
27.3.1 Facial Action Unit (FAU) Signals
27.3.2 Video Classification Model
27.3.3 Attention Module
27.4 Experiments on Video-Based Deception Detection
27.4.1 Implementation Details
27.4.2 Experiments on Real-Life Trial Dataset
27.4.3 Experiments on Bag-of-Lies
27.4.4 Experiments on the Resistance Game
27.5 Attention Visualization
27.6 Conclusion
References
28 Manufacturing the Future via Dynamic Data Driven Applications Systems (DDDAS)
28.1 Introduction
28.2 Case Study: PM-Scheduling in a Semiconductor Die Fab
28.3 DDDAS for Manufacturing the Future
28.3.1 Nanomanufacturing
28.3.2 Modeling of Materials Design
28.3.3 Biomanufacturing
28.3.4 Advanced Robotics and e-Manufacturing
28.4 Conclusions
References
29 DDDAS in the Social Sciences
29.1 Introduction
29.2 DDDAS and Social Simulations
29.3 AIMSS
29.3.1 High-Level Architecture
29.3.1.1 Data Selection
29.3.1.2 Simulation Adjustment
29.3.2 The AIMSS Scientific Workflow
29.3.3 State-of-the-Art Techniques
29.4 A Social Science Case Study
29.4.1 Data Generation and Interpretation
29.4.2 Generalization Using Data Mining
29.4.3 Consistency Checking
29.4.3.1 Formalizing Rules
29.4.3.2 Incremental Consistency Checking
29.5 Summary
29.5.1 Autonomous Learning and Model Adaptation
29.5.2 Multiple Ontologies
29.5.3 Data Selection
References
Part XI Cybersystems – CyberSecurity
30 Anomaly-Detection Defense Against Test-Time Evasion Attacks on Robust DNNs
30.1 Introduction and Overview
30.2 Background on TTE Attacks and ADA
30.3 Background on Robust DNNs
30.3.1 Key Limitations of Robust Classification Defense Against TTE Attacks
30.4 Experimental Results
30.4.1 Summary of Experimental Setting
30.4.2 Evaluation Metrics
30.4.3 Detection of Adversarial Images with Varying Attack Strength
30.4.4 Results on Conventional DNN, Weakly and Strongly Robust Trained DNN
30.4.5 TTEs with Increasing Attack Strength
30.5 Conclusions and Future Work
References
31 Dynamic Data-Driven Approach for Cyber-Resilient and Secure Critical Energy Systems
31.1 Introduction
31.2 Related Work
31.3 Enhancing Power Grid Cyber Security and Resilience Using Software-Defined Networking
31.4 High-Fidelity Testbed Combining Simulation and Emulation Systems for Power Grid Planning and Evaluation
31.5 SDN-Enabled Campus Microgrid Deployment
31.6 Application: Self-Healing PMU Network
31.6.1 Self-Healing PMU Network Architecture
31.6.2 Optimization Model and Formulation
31.6.3 Evaluation
31.7 Concluding Remarks
References
Part XII Design-Computer Systems
32 Dynamic Network-Centric Multi-cloud Platform for Real-Time and Data-Intensive Science Workflows
32.1 Introduction
32.2 Application Domain Challenges
32.2.1 Collaborative Adaptive Sensing of the Atmosphere (CASA)
32.2.1.1 Nowcast
32.2.1.2 Wind Speed
32.2.1.3 Hail
32.2.1.4 Contouring
32.3 DyNamo System Description
32.3.1 Integrated, Multi-cloud Resource Provisioning
32.3.2 Dynamic Data Movement Across Infrastructures and External Repositories
32.3.3 Mobius: DyNamo's Network-Centric Provisioning Platform
32.3.4 Workflow Automation
32.3.4.1 CASA Pegasus Nowcast Workflow
32.3.4.2 CASA Pegasus Wind Speed Workflow
32.3.4.3 CASA Pegasus Hail Workflow
32.3.5 Workflow and Data Visualization
32.4 Evaluation of DyNamo System
32.4.1 CASA Testcases
32.4.2 Experimental Infrastructure
32.4.3 Experimental Results
32.4.3.1 Effect of Cluster Parallelism in the Nowcast Workflow
32.4.3.2 Performance Study of Nowcast Workflows
32.4.3.3 Improving Data Movement Performance
32.4.3.4 Compute Slot Utilization
32.5 CASA's Operational Experience with DyNamo
32.5.1 Real-Time Operations
32.5.2 Simulated Playback of a Severe Weather Case Study
32.6 Related Work
32.7 Conclusion
References
33 INDICES: Applying DDDAS Principles for Performance Interference-aware Cloud-to-Fog Application Migration
33.1 Introduction
33.2 System Model and Assumptions
33.2.1 Components of Application Response Times
33.2.2 Architectural Model
33.2.3 Application Model
33.2.3.1 Application Performance
33.2.3.2 Migration Cost
33.3 Problem Statement and its Formulation
33.3.1 Performance Estimation Challenges
33.3.1.1 Workload Estimation
33.3.1.2 Hardware Heterogeneity
33.3.1.3 Performance Interference
33.3.2 Optimization Problem Formulation
33.3.2.1 Objective Function
33.3.2.2 Optimization and Constraints
33.4 Design of the DDDAS-based INDICES Framework
33.4.1 INDICES Architecture and Implementation
33.4.2 Execution Time Estimation via DDDAS Model Execution
33.4.2.1 Offline Phase
33.4.2.2 Online Phase
33.4.3 Network Latency Estimation
33.4.4 State Transfer Estimation
33.4.5 Solving the Optimization Problem at Runtime
33.5 Experimental Validation
33.5.1 Experimental Setup
33.5.2 DDDAS Target Application Use Case
33.5.3 Evaluating the Performance Estimation Model
33.5.4 Evaluating the Server Selection Algorithm
33.6 Related Work
33.6.1 Network Latency-based Server Selection
33.6.2 Performance Interference-aware Server Selection
33.6.3 Performance-aware Edge Computing
33.7 Conclusions
References
34 Adaptive Routing for Hybrid Photonic–Plasmonic (HyPPI) Interconnection Network for Manycore Processors Using DDDAS on the Chip
34.1 Introduction
34.2 Related Work
34.3 Weighted Recent Communication Driven (WRECD) Adaptive Routing Algorithm
34.3.1 Data Collection
34.3.2 Next-Hop Selection
34.3.3 Reconfiguration
34.4 Hybrid Photonic–Plasmonic Interconnect (HYPPI)
34.5 Router Design
34.6 Integrating WReCD Design
34.6.1 Performance of Express Links Based on Different Technologies
34.6.2 Integrating WReCD Design with HyPPI Links
34.7 Results
34.7.1 Simulation Setup
34.7.2 Simulation Results on RealTraffic Patterns
34.7.2.1 Latency
34.7.2.2 Energy
34.8 Conclusion
References
Index