Program PI Meeting Date: Sept 30 Oct 2, 2013 East/West Falls Church - - PowerPoint PPT Presentation

1

Distribution A: Approved for Public Release, Unlimited Distribution

Integrity  Service  Excellence

Frederica Darema, Ph. D., IEEE Fellow AFOSR Air Force Research Laboratory

Dynamic Data Driven Applications Systems (DDDAS) Program PI Meeting

Date: Sept 30 – Oct 2, 2013

East/West Falls Church Room,VTech/BRICC 900 North Glebe Road, Arlington, VA 22203

2

Distribution A: Approved for Public Release, Unlimited Distribution

DDDAS Program PI Meeting Overview

• The Objectives of the Meeting:
• overview of the status of the projects
• highlight accomplishments
• explore opportunities for cross-leverage of progress

across projects

• explore opportunities for cross-leverage of other US

and international programmatic activities

• discuss directions and opportunities for the future
• Agenda:
• 3-day meeting
• individual project presentations with Q&A
• plenary discussions on groups of projects

3

Distribution A: Approved for Public Release, Unlimited Distribution

Dynamic Data Driven Applications Systems (DDDAS)

• F. Darema

Experiment Measurements Field-Data (on-line/archival) User

Dynamic

Feedback & Control

Loop DDDAS: ability to dynamically incorporate additional data into an executing application, and in reverse, ability of an application to dynamically steer the measurement process

Measureme ment nts Exper erime ment nts Field-Dat ata User

“revolutionary” concept enabling design, build, manage, understand complex systems

InfoSymbiotic Systems

Dynamic Integration of Computation & Measurements/Data Unification of Computing Platforms & Sensors/Instruments (from the High-End to the Real-Time,to the PDA)

DDDAS – architecting & adaptive mngmnt of sensor systems

Challenges: Application Simulations Methods Algorithmic Stability Measurement/Instrumentation Methods Computing Systems Software Support

Synergistic, Multidisciplinary Research

4

Distribution A: Approved for Public Release, Unlimited Distribution

DDDAS - Clearly articulated concept/paradigm:

• integration of application simulation/models with the application

instrumentation components in a dynamic feed-back control loop

• speedup of the simulation, by replacing computation with data in specific

parts of the phase-space of the application and/or

• augment model with actual data to improve accuracy of the model, improve

analysis/prediction capabilities of application models

• enable ~decision-support capabilities w simulation-modeling accuracy
• dynamically manage/schedule/architect heterogeneous resources, such as:
• networks of heterogeneous sensors, or networks of heterogeneous controllers

increased computation/communication capabilities; ubiquitous heterogeneous sensing

• unification from the high-end to the real-time data acquisition and control

 DDDAS is more powerful and broader paradigm than Cyber-Physical Systems

Advances in Capabilities through DDDAS

5

Distribution A: Approved for Public Release, Unlimited Distribution

DDDAS/AFOSR BAA and Technology Horizons

• Context of Key Strategic Approaches of the Program

– Multidisciplinary Research – Focus of advancing capabilities along the Key Areas identified in the Technology Horizons, and the Energy Horizons and Global Horizons Reports

• Autonomous systems
• Autonomous reasoning and learning
• Resilient autonomy
• Complex adaptive systems
• V&V for complex adaptive systems
• Collaborative/cooperative control
• Autonomous mission planning
• Cold-atom INS
• Chip-scale atomic clocks
• Ad hoc networks
• Polymorphic networks
• Agile networks
• Laser communications
• Frequency-agile RF systems
• Spectral mutability
• Dynamic spectrum access
• Quantum key distribution
• Multi-scale simulation technologies
• Coupled multi-physics simulations
• Embedded diagnostics
• Decision support tools
• Automated software generation
• Sensor-based processing
• Behavior prediction and anticipation
• Cognitive modeling
• Cognitive performance augmentation
• Human-machine interfaces

Top KTAs identified in the 2010 Technology Horizons Report

• Autonomous systems
• Autonomous reasoning and learning
• Resilient autonomy
• Complex adaptive systems
• V&V for complex adaptive systems
• Collaborative/cooperative control
• Autonomous mission planning
• Cold-atom INS
• Chip-scale atomic clocks
• Ad hoc networks
• Polymorphic networks
• Agile networks
• Laser communications
• Frequency-agile RF systems
• Spectral mutability
• Dynamic spectrum access
• Quantum key distribution
• Multi-scale simulation technologies
• Coupled multi-physics simulations
• Embedded diagnostics
• Decision support tools
• Automated software generation
• Sensor-based processing
• Behavior prediction and anticipation
• Cognitive modeling
• Cognitive performance augmentation
• Human-machine interfaces

DDDAS … key concept in many of the objectives set in Technology Horizons

6

Distribution A: Approved for Public Release, Unlimited Distribution

AFOSR DDDAS Program

DDDAS Program Research Components

• Application Modeling/Simulation
• Application Algorithms
• Systems Software
• Instrumentation methods
• Program announced in AFSOR BAA-2011(posted in Spring2011)
• Projects awarded in FY11, FY12, FY13
• Additional proposals have been submitted and are reviewed

for FY14 funding;

• Additional WP and proposals are expected to by submitted

inFY14

7

Distribution A: Approved for Public Release, Unlimited Distribution

PI Meeting Agenda Day 1 – September 30, 2013

Day 1 – Morning Session

8:00am-8:30am – Introduction to the Program - Overview of PI Meeting-- Darema 8:30am-10:30am Materials modeling

• Development of a Stochastic Dynamic Data-Driven System for Prediction of Materials Damage

– PI: Tinsley Oden (UT Austin), and Team

• Developing Data-Driven Protocols to study Complex Systems: The case of Engineered Granular

Crystals (EGC) – PI: Yannis Kevrekidis (Princeton Univ) , and Team

• Dynamic Data-Driven Modeling of Uncertainties and 3D Effects of Porous Shape Memory Alloys

– PI: Craig Douglas (U of Wyoming), and Team

• Dynamic, Data-Driven Modeling of Nanoparticle Self Assembly Processes

–

• Y. Ding (TAMU),and Team

10:30am-10:45am --Break 10:45am-12:15pm AirVehicle Structural HealthMonitoring – Environment Cognizant

• Advanced Simulation, Optimization, and Health Monitoring of Large Scale Structural Systems

– PI: Yuri Bazilevs (UCSD), and Team

• Dynamic Data-Driven Methods for Self-Aware Aerospace Vehicles

– PI: Karen Willcox (MIT), and Team

• Bayesian Computational Sensor Networks for Aircraft Structural Health Monitoring

– PI: Thomas Henderson (U. of Utah)

12:15pm-1:00pm --Lunch

8

Distribution A: Approved for Public Release, Unlimited Distribution

PI Meeting Agenda Day 1 – September 30, 2013

Day 1 – Afternoon Session

1:00pm -2:00pm Energy Efficiencies Energy-Aware Aerial Systems for Persistent Sampling and Surveillance

– PI: Erik Frew (U of Colorado-Boulder), and Team

• DDDAMS-based Real-time Assessment and Control of Electric-Microgrids

– PI: Nurcin Celik (University of Miami)

2:00pm -3:00pmSpatial Situational Awareness (UAV Swarms + Ground Systems Coordination)

• Application of DDDAS Principles to Command, Control and Mission Planning for UAV Swarms

– PI: Greg Madey (U. Of Notre Dame), and Team

• DDDAMS-based Urban Surveillance and Crowd Control via UAVs and UGVs

– PI: Young-Jun Son ( University of Arizona),, and Team

3:00pm -3:15pm --Break 3:15pm -4:45pm (UAV Swarms + Ground Systems Coordination) –New Starts (15mins each)

• Dynamic Systems for Individual Tracking via Heterogeneous Information Integration and Crowd Source

Distributed Simulation – PI: Richard Fujimoto (Georgia Tech), and Team

• An Integrated approach to the Space Situational Awareness Problem

– PI: Suman Chakravorty (TAMU) , and Team

• A Dynamic Data Driven Cognitive Control Architecture for Exploration

– PI: Jose Principe (U. of Florida), and Team

4:45pm -5:30pm – Discussion of all projects discussed in Day 1

9

Distribution A: Approved for Public Release, Unlimited Distribution

PI Meeting Agenda Day 2 – September 30, 2013

Day 2 – Morning Session

8:15am-10:00am – Spatial Situational Awareness (Co-operative Sensing UAV-Ground-Space)

• DDDAS for Object Tracking in Complex and Dynamic Environments (DOTCODE)

– PI: Anthony Vodacek (RIT) and Team

• Dynamic Data Driven Adaptation via Embedded Software Agents for Border Control Scenario

– PI: Shashi Phoha (Penn State), and Team

• Multiscale Analysis of Multimodal Imagery for Cooperative Sensing

– PIs: Erik Blasch, Guna Seetharaman, RI Directorate, AFRL

• New Globally Convex Models for Vision Problems using Variational Methods (LRIR)

– PI: Guna Sheetharanam, AFRL-RI

10:00am-10:15am –Break 10:15am-11:45am – Spatial Situational Awareness (Co-operative Sensing UAV-Ground-Space) – cont’d

• Adaptive Stream Mining: A Novel Dynamic Computing Paradigm for Knowledge Extraction

– PI: Shuvra Bhattacharyya (U. Of Maryland) and Team

• Stochastic Logical Reasoning for Autonomous Mission Planning

– PI: Carlos A. Varela (RPI)

• Hybird Systems Modeling and Middleware-enabled DDDAS for Next-generation US Air Force Systems

– PI: Aniruddha Gokhale (Vanderbilt U.), and Team (Doug Schmidt)

11:45am-1:00pm --Lunch

10

Distribution A: Approved for Public Release, Unlimited Distribution

PI Meeting Agenda Day 2 – September 30, 2013

Day 2 – Afternoon Session

1:00pm -3:00pm Space Weather and Atmospheric Events – Modeling/Observations

• Transformative Advances in DDDAS with Application to Space Weather Modeling

– PI: Dennis Bernstein (U. of Michigan), and Team

• DDDAS Approach To Volcanic Ash Transport & Dispersal Forecast

– PI: Abani Patra (Univ at Buffalo), and Team

• Fluid SLAM and the Robotic Reconstruction of Localized Atmospheric Phenomena

– PI: Sai Ravela (MIT)

• A Framework for Quantifying and Reducing Uncertainty in InfoSymbiotic Systems Arising in

Atmospheric Environments – PI: Adrian Sandu (Virginia Tech )

3:00pm -3:15pm --Break 3:15pm -4:15pm Space Weather&AtmosphEvents – Mod/Obs –New Starts (15mins each)

• Dynamic Data-Driven UAV Network for Plume Characterization

– PI: Kamran Mohseni (U. of Florida)

• Optimized Routing of Intelligent,Mobile Sensors for Dynamic, Data-Driven Sampling

– PI: Derek Paley (UMD)

4:15pm -5:30pm – Discussion of all projects discussed in Day 2 (& Day 1)

11

Distribution A: Approved for Public Release, Unlimited Distribution

PI Meeting Agenda Day 3 – September 30, 2013

Day 3 – Morning Session 8:00am-10:00am Systems Software

• An Adaptive Property-Aware HW/SW Framework for DDDAS

– PI: Philip Jones (Iowa State U.), and Team

• Amorphous Polyhedral Model for Stochastic Control of Autonomous UAVs

– PI: Sanjay Rajopadhye (Colorado State)

• PREDICT: Privacy and Security Enhancing Dynamic Information Collection and Monitoring

– PI: Vaidy Sunderam (Emory U.), and Team

• DDDAS-based Resilient Cyberspace (DRCS)

– PI: Salim Hariri (University of Arizona. Tucson), and Team

10:00am-10:15am –Break 10:15am-11:00am – Systems Software (cont’d)

• Xiaoming Li…
• Cloud-Based Perception and Control of Sensor Nets and Robot Swarms (New Start – 15mins)

– PI: Geoffrey Fox (Indiana U.), and Team 11:15pm -12:00noon – Discussion of all projects discussed in Morning of Day 3

• 12noon-1pm – Lunch – (Working Lunch?)

Day 3 – Afternoon Session

• 1pm-2:30pm Discussion of all projects – Collaborations, Directions in the Program

12

Distribution A: Approved for Public Release, Unlimited Distribution

PI Meeting Agenda

Some of the Discussion items for Day 3 – September 30, 2013

DDDAS Outreach and Repositories of DDDAS work

Presently:

• www.dddas.org

– contains reports of funding agencies sponsored workshops – slides and papers of community organized DDDAS Workshops

• DDDAS Workshop at ICCS (10-year history)
• Other DDDAS workshops organized by the community (examples below)
• Papers in ICCS Proceedings of the DDDAS Workshop
• Papers published by the research community

Other Outreach activities:

• Bridge with other funding agencies in the US, EU + OtherEurope, Asia Outreach

to additional research communities; e.g.:

– Darema: DDDAS Panel at 2013 American Controls Conference (June 17, 2013) – Dennis Bernstein: DDDAS Workshop at 2014 American Controls Conference – Ana Cortes, et al: DDDAS Workshop on Fire Modeling - EU-US-Asia – Craig Douglas, Abani Patra, and Raj Buyya(AUS), ICCS2014/DDDAS-Workshop, AUS

• Include in www.dddas.org more systematically pointers to all DDDAS papers

published by the research community

• Complete book on DDDAS – chapters representing projects

– uniform conceptual format of chapters; not a compendium of papers – effort has started; will update update/add & complete; set a timeline

{Need to update website format and look}

13

Distribution A: Approved for Public Release, Unlimited Distribution

• Application modeling (in the context of dynamic data inputs)
• dynamically invoke/select appropriate application components (models/algorithms)

depending on streamed data

• multi-modal, multi-scale – dynamically invoke multiple scales/modalities
• dynamic hierarchical decomposition (computational platform - sensor) and partitioning
• interfacing applications with measurement systems
• Algorithms
• tolerant to perturbations of dynamic data inputs
• UQ, uncertainty propagation
• Measurements
• multiple modalities, space/time-distributed
• heterogeneous data management
• Systems supporting dynamic runtime environments
• extended spectrum of platforms
• - beyond traditional computational grids, beyond the “traditional” cloud,

to include sensor/instrumentation grids

• dynamic execution support on heterogeneous environments

Fundamental Science and Technology Challenges for Enabling DDDAS Capabilities

14

Distribution A: Approved for Public Release, Unlimited Distribution

Topics for discussion at the End of the Day

Please upload your presentations on the laptop at the podium

End of Day 1

• Multi-model selection; Dynamic Model selection <-> DDDAS (e.g. Tinsley) and selection across

multiple-measurements – dynamic selection of measurements for model bias (e.g.: Ding); and both / multi-fidelity simulations/sensing (e.g. Wilcox; Celik; Son)

• Improve accuracy / derive models with actual data (e.g. Kevrekidis; Bernstein)
• VVUQ (Henderson); Bayesian methods (several talks…)
• Interactions w industry (e.g. Douglas; Wilcox; Henderson/Boeing; Blasch; Varela/Wilcox/Abani – IBM;

Ravela)

• Interactions with other Agencies (e.g. Terry/NSF; Bazilevs <-> DOE)
• Interactions w other parts of the AF (Celik; Madey; Chakravorty; Hariri) --- LabTasks

End of Day 2

• Cross-leverage across sub-groups of projects modeling methods/algorithms/software frameworks

(e.g.: Day1morning&Day2/2ndpart of morning; Day1/afternoon; Day2/1stpart of morning – Situational Awareness;

• + Day1 add-ons

End of Day 3

• Programming Models/Environments for DDDAS
• xxx
• + Day1 & Day2 add-ons

15

Distribution A: Approved for Public Release, Unlimited Distribution

Break-out Groups - Discussions

Please upload your presentations on the laptop at the podium

End of Day 2

• BG1 - Structural Health monitoring/assessment/control;
• BG2 – UAVs coordination - autonomous &optimized operation and mission execution
• BG3 – Surveillance
• Environmental & Space

BGs 2min Report

• ….

End of Day 3

• BG1 - Core Methodologies - I

– Application Tools (Models, Algorithms, supporting tools) – Abani – “dynamicity of data”; DDDAS presence at AAIA

• BG2 - Core Methodologies - II

– Systems sw – architectural sw frameworks infrastructure – Varela – data streams; data consistency; domain specific languages which can hide non-functional concerns; how to reason about DDDAS apps, hide inputs/outputs; challenge of the infrastructure – need software frameworks – dynamic workflows.

• BG3 - DDDAS & UAVs/Mobile Sensing

– Madey –

• Collaboration opportunities GTRI w WalterRobbins AFB (possible UAV station) - Fujimoto
• Other Agencies ~ mobile sensing – Madey has conversations with NRO; SPAWAR – Dahlgreen;
• SBIRs (phse I/II)
• UAVs – AR large UAVS; Army – small; Swarms…
• + 3 slides

16

Distribution A: Approved for Public Release, Unlimited Distribution

Conceptual format of the DDDAS Springer Book Chapters

Book Outline: Sections of overarching areas covered by DDDAS projects Each Section will contain several chapters (~projects) The book will include AFOSR projects and other (e.g. previously NSF funded )

• L1: The scientific area addressed by the project – and how DDDAS is used to address

challenges / create new capabilities. The first part of the chapter should be at a level such that a person with scientific/technical background can understand it, and learn from it, (“Scientific American” article- level). The Introduction should be at that level.

• L2: The second level (subsection 2, in the chapter) which explain s(in more technical

detail than the overview) the state of the art, the challenges and methods the work pursues, and how specifically DDDAS is used to address the challenges (in more specific terms than in the overview).

• L3: the next set of subsections in your chapter would contain your specific approaches,

as you would present them in a journal paper. For example, like the papers submitted for the ICCS workshops and earlier versions for the book (Kluwer)/Springer, updated. ICCS DDDAS papers, over the years the papers have addressed different facets of the

• ngoing work in a project. The contribution in the book should present an integrated

synthesis of these efforts/advances.

• L4: is the summary subsection of the chapter. It should include discussion of the impact
• f your individual methods, the specific impact in using the DDDAS concept in your

results, new capabilities enabled, new venues and directions created in your field, and possibly how they may apply to other fields. This section targeted for a range of possible readers, from experts in the area, to those with scientific/technical backgrounds in related or other areas, and senior and graduate students, … who can learn from the new methods presented and be inspired to utilize in their own work.