Visible to the public File preview

Adap%ve
 Vehicle
 Make
 (AVM)
 

  Mr.
 Paul
 Eremenko,
 Program
 Manager
  LTC
 Nathan
 Wiedenman,
 Deputy
 Program
 Manager
 
  TacHcal
 Technology
 Office
 
 
 
  August
 1,
 2011
 
 
 

APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
 

1
 

Historical
 schedule
 trends
 with
 complexity
 
240 220 ) s h 200 t n o m ( 180 g n 160 i t s e T140 d n a 120 , n o i 100 t a r g e 80 t n I , 60 n g i s e 40 D 20

Next-Gen Platform New IC design flow

Historical Cost Growth (not adjusted for inflation) Aerospace Systems (1960–present) 8-12%/yr Automobiles (1960–present) Integrated Circuits (1970–present) 4%/yr ~0%/yr

MIL-STD-499A

Aerospace Vehicle 1990s

~5X Reduction in Development Effort

New automotive design flow Automobile 1960s Aerospace Vehicle 1960s Automobile 1990s Pentium Integrated Circuit 1960s Intel 8088 Intel 286 Intel 386 Automobile Next Gen

META Goal
Integrated Circuit Next Gen Xeon

?
 

0 1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+09

1.E+10

Complexity* [Part Count + Source Lines of Code (SLOC)]

APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
 

2
 

Status
 quo
 approach
 to
 managing
 complexity
 
SWaP
 used
 as
 a
 proxy
  metric
 f or
 cost,
 and
 dis-­‐ incentivizes
 abstraction
  in
 design

System
 decomposed
  based
 on
 arbitrary
  cleavage
 lines
 .
 .
 .

MIL-­‐STD-­‐499A
 (1969)
 systems
 engineering
  process:
 as
 employed
 today

Conventional
 V &V
 techniques
  do
 not
 scale
 to
 highly
 complex
  or
 adaptable
 systems–with
  large
 or
 infinite
 numbers
 of
  possible
 states/configurations

Re-Design
Cost Optimization

System Functional Specification

System Layout

Verification & Validation

SWaP Optimization

...
Power Data & Control Thermal Mgmt

Subsystem Design

Subsystem Testing

Resulting
  architectures are
 fragile
  point
 designs

SWaP Optimization

...

Component Design

Component Testing

.
 .
 .
 and
 detailed
 design
  occurs
 within
 these
  functional
 stovepipes
SWaP = Size, Weight, and Power V&V = Verification & Validation

Unmodeled
 and
 undesired
  interactions
 lead
 to
 emergent
  behaviors
 during
 integration
Desirable interactions (data, power, forces & torques) Undesirable interactions (thermal, vibrations, EMI)

APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
 

3
 

The
 technical
 problem
 is
 in
 the
 seams
 
Between
 stages
 of
 produc%on
 →
 


 
 Between
 system
 components
 

Source:
 MIT
 ESD
 (deWeck
 et
 al.,
 2008)
 

Between
 people
 &
 organiza%ons
 →
 
Source:
 
 DDR&E/SE
 (Flowe
 et
 al.,
 2009)
 

APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
 

4
 

AdapHve
 Vehicle
 Make
 vision
 
Shorten
 development
 %mes
 for
 complex
 defense
 systems
 [META]
 
•  •  •  •  Raise
 level
 of
 abstracHon
 in
 design
 of
 cyber-­‐electromechanical
 systems
  Enable
 correct-­‐by-­‐construcHon
 designs
 through
 model-­‐based
 verificaHon
  Compose
 designs
 from
 component
 model
 library
 that
 characterizes
 the
 “seams”
  Rapid
 requirements
 trade-­‐offs;
 opHmize
 for
 complexity
 &
 adaptability,
 not
 SWaP
 


 

ShiI
 product
 value
 chain
 toward
 high-­‐value
 design
 ac%vi%es
 [iFAB]
 


 

•  Bitstream-­‐configurable
 foundry-­‐like
 manufacturing
 capability
 for
 defense
 systems
  •  Rapid
 switch-­‐over
 between
 designs
 with
 minimal
 learning
 curve
  •  “Mass
 customizaHon”
 across
 product
 variants
 and
 families
  •  Crowd-­‐sourcing
 infrastructure
 to
 enable
 open-­‐source
 development
 of
 cyber-­‐ electromechanical
 systems
 [vehicleforge.mil]
  •  Prize-­‐based
 AdapHve
 Make
 Challenges
 culminaHng
 in
 an
 Infantry
 FighHng
 Vehicle
 for
  tesHng
 alongside
 a
 program
 of
 record
 [FANG]
  •  MoHvate
 a
 new
 generaHon
 of
 designers
 and
 manufacturing
 innovators
 [MENTOR]
 

Democra%ze
 design
 [FANG]
 


 

APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
 

5
 

META
 

APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
 

6
 

NoHonal
 iFAB
 foundry
 configuraHon
 
Paint & Finish

Assumptions: •  100k ft2 total space. •  Need not be geographically co-located. Additive/Subtractive •  All custom components in-sourced. Manufacturing •  All unmod COTS components out-sourced.
Sheet Metal Fabrication Fuels & Tribology Composites Tape Laying CNC Brake 3D Printer Laser Cutter Automated Harness Loom Fuel Cell Test Set CNC CMM Articulating CMM Autoclave

•  Drill & fill •  Wire bundles •  Robotic assembly

Welding

Harness Buildup

Electronics Fabrication

Paint Booth

Welding Robots Anodizing Tank

CNC Mill 5-Axis 6-Axis Robots

Assembly

Swaging Press

CNC
 Tube
  Bender
 
Dynamometer

Machine Instructions (STEP-NC, OpenPDK)

Logistics Tube Bending Hydraulics & Pneumatics QA / QC

iFAB Foundry Configuration

Product MetaRepresentation

APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
 

7
 

Crowd-­‐sourcing
 infrastructure:
 vehicleforge.mil
 
iFAB

auxiliary systems

chassis

drivetrain

Es%mated
 Size
 of
 Component
 Model
 Library
 
Assembly
  Drivetrain
  Chassis/Armor
  Other
  Total
  Unique
 Parts
 
  (upper
 limit)
  3,000
  5,000
  7,500
  15,500
  Total
 Parts
  (lower
 limit)
  8,000
  12,000
  10,000
  30,000
  Library
 Parts
 
  (unique
 x
 5)
  15,000
  25,000
  37,500
  72,500
 

Elements
 of
 a
 Component
 Model
 
Physical
 aiributes
 
•  size
 and
 shape
  •  mass
 properHes
  •  elastodynamics
  •  data
  •  power
  •  mechanical
 

Undesirable
 emissions
 
 
•  thermal
  •  electro-­‐magneHc
  •  vibraHonal
  •  blackbox
 model
  •  failure
 modes
 

Interfaces
 

Performance
 

Note:
 EsImates
 are
 at
 the
 numbered
 part
 level.
 Cables
 and
 circuit
 boards
  counted
 as
 single
 part.
 Excludes
 variable
 mission
 equipment,
 soPware.
 

APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
 

8
 

Fast,
 Adaptable
 Next-­‐generaHon
 Ground
 vehicle
 (FANG)
 
Mobility/Drivetrain
 
  Challenge
  Chassis/Integrated
 Survivability
  Challenge
  Total
 PlaYorm
  Challenge
 

SCOPE
 

SCOPE
 

SCOPE
 

PARTICIPANT
 POOL
  INCENTIVE
 
•  Global
 

•  Vehicle
 drivetrain
 to
 meet
 IFV
 speed,
  efficiency,
 terrain,
 reliability
 objecHve
  •  Available
 model
 library
 to
 include:
 
  •  Hybrid-­‐electric
 systems
  •  Novel
 ground
 interfaces
 

PARTICIPANT
 POOL
  INCENTIVE
 
•  Global
 

•  Chassis
 and
 armor
 design
 to
 meet
 principal
 IFV-­‐ like
 survivability
 objecHves
  •  Available
 model
 library
 to
 include:
 
  •  Advanced
 armor
 concepts
  •  Novel
 configs
 (monocoque,
 v-­‐hulls)
 

PARTICIPANT
 POOL
  INCENTIVE
 
•  Global
 

•  Complete
 IFV
 based
 on
 core
 Army
 objecHves
  and
 disHlled
 requirements
 

DESIGN
 AGGREGATION
  BUILD
 APPROACH
 

•  Prize
 $1M
 for
 winning
 design
  •  Winner(s)
 judged
 based
 on
 mulH-­‐objecHve
  weighHng
 funcHon
  •  Use
 of
 META
 metalanguage
 required
  •  Use
 of
 vehicleforge.mil
 opHonal
  •  iFAB
 foundry
 build
 for
 top
 design(s)
 

DESIGN
 AGGREGATION
  BUILD
 APPROACH
 

•  Prize
 $1M
 for
 winning
 design
  •  Winner(s)
 judged
 based
 on
 mulH-­‐objecHve
  weighHng
 funcHon
  •  Use
 of
 META
 metalanguage
 required
  •  Use
 of
 vehicleforge.mil
 opHonal
  •  iFAB
 foundry
 build
 for
 top
 design(s)
 

•  Prize
 $2M
  •  Winner
 judged
 based
 on
 saHsfacHon
 of
  constraints
 and
 mulH-­‐airibute
 preference
  funcHon
 (i.e.,
 enHrely
 objecHve
 approach)
 

DESIGN
 AGGREGATION
  BUILD
 APPROACH
 

•  Use
 of
 META
 metalanguage
 required
  •  Use
 of
 vehicleforge.mil
 opHonal
  •  iFAB
 foundry
 build
 for
 top
 design(s)
 

APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
 

9
 

Experimental
 Crowd-­‐derived
 Combat
 support
 Vehicle
 (XC2V)
 
Goal
•  Experiment in crowd-sourced design •  Militarily-relevant application •  Existing (simple) commercial infrastructure

Approach

•  Utilize existing social network of ~20,000 from Local Motors (increased by ~3,000) •  Crowd-source design of a combat support vehicle •  $10k in prizes •  Build in existing micro-factory

Results
•  159 final designs submitted •  100 of “high caliber” according to DARPA Service Chiefs Fellows •  4 week design period •  14 week build period
APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
  10
 

Manufacturing
 ExperimentaHon
 and
 Outreach
 (MENTOR)
 
Goal
 
•  Educate,
 moHvate,
 and
 inspire
 a
 next-­‐ generaHon
 cadre
 of
 designers
 and
  manufacturing
 innovators
  •  Inculcate
 AVM-­‐type
 design
 methods
 such
 that
  they
 become
 pervasive
 in
 subsequent
  generaHons
 of
 engineers
 

Approach
 
•  Design
 collaboraHon
 using
 modern
 CAD
 tools
  and
 convenHonal
 social
 networking
 media
  •  Distributed
 manufacturing
 across
 networks
 of
  schools
 equipped
 with
 various
 digital
  manufacturing
 equipment
  •  Run
 compeHHve
 prize
 challenges
 for
 design
  and
 build
 of
 moderately
 complex
 systems
 (e.g.
  go-­‐carts,
 mobile
 robots,
 small
 UAVs,
 etc.)
  •  Outreach
 ObjecHves:
 
•  10
 schools
 in
 CY12
  •  100
 schools
 in
 CY13
  •  1,000
 schools
 in
 CY14
 

•  ParHcipaHon
 by
 domesHc
 and
 foreign
 schools
 

Picture
 credits:
 Robot
 image
 source
 -­‐
 goroboHcs.net;
 Los
 Gatos
 HS,
  CA;
 Loy
 Norris
 HS
 ,
 MI;
 Stoney
 Creek
 HS,
 CA;
 Lakeridge
 HS,
 OR;
  New
 Smyrna
 Beach
 HS,
 FL;
 Longhill
 HS,
 West
 Sussex,
 UK;
  Brockton
 HS,
 MA
 
 

APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
 

11
 

AVM
 porpolio
 schedule
 
CY09
  FY10
  CY10
  FY11
 
META
  Tools
 &
 Metalanguage

CY11
  FY12
 

CY12
  FY13
 

CY13
  FY14
 

CY14
 

CY15
  FY15
 


 

META-­‐X
  Tools
 MaturaIon C2M2L-­‐1
  Mobility
 Drivetrain
 Library


 
 
 
 
 


 
C2M2L-­‐3
  Full
 IFV
 Model
 Library

C2M2L-­‐2
  Chassis
 &
 Surv
 Library

iFAB
  Tools
 &
 Model
 Library

XC2V
  CS
 Pilot
 

iFAB
  Configure,
 Build,
 and
 Operate
 Foundry vehicleforge.mil
  Crowd
 Source
 Design
 Infrastructure
 Development
 and
 Maintenance FANG Execute
 AdapIve
 Make
 Challenges
 


 


 

Mobility/Drivetrain
 
  Challenge
  Chassis/Integrated
 
  Survivability
 Challenge
  Total
 PlaYorm
 Challenge
 

Design
 

Fab
 

T&E
  ComparaIve
 TesIng
 

USMC/US
 Army
 
 IFV
 Prototype
  MENTOR
  Distributed
 Manufacturing
 High
 School
 Outreach
  SBIR
 PorYolio
  Large-­‐Scale
 Manufacturing
 in
 QuanIIes
 of
 One
  APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
  12
 

AdapHve
 Vehicle
 Make
 performer
 community
 
META
 
Adven%um
 Enterprises
  BAE
 Systems
  Boeing
  Dassault
 Systèmes
  IBM
 Haifa
 Research
 Lab
 
  MIT
 (Dr.
 Rhodes)
  MIT
 (Prof.
 Wilcox)
  Rockwell
 Collins
  Smart
 Info
 Flow
 Tech
 (SIFT)
  SRI
 Interna%onal
  United
 Tech
 Research
 Ctr
  Vanderbilt
 Univ
 (Dr.
 Bapty)
  Vanderbilt
 Univ
 (Dr.
 Neema)
  Xerox
 PARC
  Formal
 metalanguage
 capable
 of
 integraHng
 models
  across
 mulHple
 abstracHon
 levels
 and
 perspecHves
  MulH-­‐abstracHon-­‐level
 design
 framework
 with
  simulaHon
 traces
 for
 cerHficaHon
 
  Metrics
 suite
 to
 support
 mulHple
 disciplines
 using
 rich
  corporate
 historical
 data
 
  Extension
 of
 commercial
 CATIA/DELMIA
 PLM
 suite
 to
  enable
 formal
 verificaHon
  Formal
 contract-­‐based
 language
 to
 enable
 true
 mulH-­‐ domain,
 plaporm-­‐based
 design
  Epoch-­‐based
 real-­‐opHon
 theory
 to
 assess
  changeability
 costs
 in
 system
 designs
  Entropy-­‐based
 sensiHvity
 and
 variance-­‐based
  allocaHon
 theories
 for
 early
 complexity
 assessment
  Pre-­‐verified,
 reusable
 design
 paierns
 for
 hardware/ soqware
 co-­‐design
  Unified
 probabilisHc
 and
 non-­‐probabilisHc
 verificaHon
  with
 counter-­‐examples
 to
 guide
 design
  Verifying
 and
 guiding
 design
 across
 abstracHon
 levels
  and
 domains
 using
 composiHonal
 framework
  Plaporm-­‐based
 generaHon
 of
 design
 space
 of
 feasible
  architectures
 with
 metric-­‐based
 selecHon
  ComposiHonal
 cross-­‐domain
 tool-­‐chain
 analysis
  templates
 that
 support
 deep
 domain
 analysis
  Rich
 model-­‐based
 approaches
 developed
 for
 soqware
  and
 VLSI
 into
 the
 CPS
 world
  FuncHon-­‐based
 framework
 for
 co-­‐verificaHon
  assessment
 and
 reasoning
 at
 early
 stages
 of
 design
  Boeing/General
 Motors
  Carnegie
 Mellon
 Univ
  Inten%onal
 SoIware
  Penn
 State
 ARL
  Univ
 of
 Delaware
  Xerox
 PARC
 

iFAB
 
Manufacturing
 capability
 and
 process
 model
 library
 with
  describing
 foundry
 resources
 &
 human
 actors
  Distributed
 agents/process
 model
 approach
 for
 two-­‐way
  interface
 between
 CAD
 and
 CAM
 systems
  Formal
 “meta
 meta”
 language
 to
 enable
 mulH-­‐domain
  co-­‐design
 of
 arHfact
 &
 manufacturing
  Agent-­‐based
 foundry
 configuraHon
 and
 trade
 space
  visualizaHon
  Developing
 composiHonal
 cross-­‐domain
 tool-­‐chain
  analysis
 templates
 for
 composites
 manufacturing
  Rapid
 construcHon
 and
 search
 of
 feasible
  manufacturability
 spaces
 and
 metrics
 for
 such
 spaces
  CreaHng
 adaptable
 soqware
 libraries
 of
 manufacturing
  processes
 perHnent
 to
 the
 fabricaHon
 of
 electro-­‐ mechanical
 components
 and/or
 assemblies
 

Georgia
 Tech
 GTRC
 

vehicleforge.mil
 
GE Research/MIT* Custom collaboration site linking to MIT DOME model repository and social network challenge platform Collaboration site based on open source distributed version control system; teamed with RedHat Collaboration site derived from KForge software and information forge site platform Credentialing users and contributions utilizing reputation-based quantitative trust management

Georgia Tech GTRI*

Vanderbilt University*

Univ of Pennsylvania*

APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
 

13
 

APPROVED
 FOR
 PUBLIC
 RELEASE.
 DISTRIBUTION
 UNLIMITED.
 

14
  1