Presentation for OMG Systems Engineering Domain Special Interest

Скачать презентацию Presentation for OMG Systems Engineering Domain Special Interest

ee1dc7336637a3a0076dc1cf4bda69d2.ppt

Количество слайдов: 41

Presentation for OMG Systems Engineering Domain Special Interest Group (SE DSIG) December 7, 2010 - Santa Clara, CA Systems Engineering Research Center (SERC) Overview and SERC RT 21 Effort: Verification, Validation, and Accreditation Shortfalls for Modeling and Simulation Overview of GIT Aspects: A Sys. ML Model-Based Approach for M&S VV&A Russell Peak (PI) Georgia Institute of Technology Model-Based Systems Engineering Center www. mbse. gatech. edu All material is copyrighted © by Georgia Tech unless otherwise noted. Permission to use for non-commercial purposes (including internal industry usage) is hereby granted only if a proper citation is given.

Contents • SERC Overview – Highlights from Art Pyster presentation at Annual SERC Research Review (ASRR) Nov 2010 - www. sercuarc. org • SERC RT 21 Overview – GIT Effort – See also RT 21 presentation at ASRR for additional RT 21 material: Do. D motivation/context, UA Huntsville AADL effort 2

Excerpts included here. - See SERC ASRR 2010 website for full version - www. sercuarc. org Systems Engineering Initiatives for Verification, Validation and Accreditation of Do. D Models and Simulations Philomena M. Zimmerman Deputy Director, Modeling, Simulation & Analysis ODDR&E/Systems Engineering/Systems Analysis Systems Engineering Research Center November 9, 2010 SERC Annual Review Nov 2010 Page-3

Acquisition Community-led VV&A High Level Task (HLT) Summaries V-AQ-2: “Risk Based Methodology for Verification, Validation and Accreditation (VV&A)” The degree of VV&A required is explicitly tied to both M&S use and the user risk incurred if the M&S does not provide accurate results. A methodology that tailors VV&A planning and implementation based on known risk factors will provide a framework in which VV&A implementation trade-offs can be made, information/fidelity requirements can be assessed, and a VV&A cost model can be developed. V-C-2: “Improving VV&A Implementation” Increase VV&A implementation and enhance M&S credibility by transforming VV&A practices from current subjective methods into objective examples or use cases. Explore emerging technologies, standards, and applicable methods that could be applied to reduce costs, schedule, and improve reuse. SERC Annual Review Nov 2010 Page-4

Improved Decision Support Quality through a Balanced Approach VV&A Strategy SERC RT 21 efforts: - GIT: Sys. ML-based approach - UA Huntsville: AADL approach SERC Annual Review Nov 2010 Page-5

Contents: SERC RT 21 – GIT Effort • Primary Content – Project VV&A objectives – Process being used – Products being produced – Progress update • Summary • Additional Background Material 6

Project VV&A Objectives – GIT Focus per updated scope 2010 -07 -20 • Primary objective – Demonstrate how to address VV&A gaps by applying Sys. ML and MBSE technology – Show in particular how VV&A can be more embedded and automated throughout the system lifecycle • Supporting sub-objectives (via “quick-look” approach) – Apply known modeling & simulation (M&S) patterns and develop new patterns where needed – Demonstrate approach by extending existing testbeds and examples (excavator testbed – next slide, other examples, . . . ) – Provide basis for developing future Do. D-specific testbeds • Terminology – Sys. ML is the Systems Modeling Language (www. omgsysml. org), which has been called “the new global language of 350 K+ systems engineers” (amazon. com) – MBSE is model-based systems engineering (vs. document-centric approach) 3/19/2018 7 7

The 4 Pillars of Sys. ML Automotive Anti-Lock Braking System Example 1. Structure 2. Behavior interaction state machine activity/ function definition use 3. Requirements Copyright © Georgia Tech and Inter. CAX. All Rights Reserved. 4. Parametrics Sys. ML and MBSE: A Quick-Start Course 8

GIT Project Team – RT 21 • Research Professionals – Selcuk Cimtalay, Ph. D – Russell Peak, Ph. D (PI) – Andy Scott – Miyako Wilson • Undergraduate Research Assistants – Brian Aikens – Drew Martin 9

Process Being Used (p 1/2) Guiding Philosophy • Enabling bottom-up/top-down hybrid approach – Iterative ubiquitous VV&A; building block VV&A – Software V&V techniques applied to systems (continuous integration/builds, junit, . . . ) • Analogies: (a) Making pizza. . . (b) “For want of a (VV&A’ed) nail. . . ” http: //en. wikipedia. org/wiki/For_Want_of_a_Nail_(proverb) 11

Process Being Used (p 2/2) GIT RT 21 Project Plan • Leverage existing examples – Illustrate technical approach in quick-look fashion – Add VV&A-oriented extensions where needed • Demonstrate sample VV&A use cases along multiple system dimensions: – system levels, tools, methods, lifecycle phases, . . . • See next slides for Project Plan specifics 12

Products Being Produced • Deliverable D 01 - Presentation & Live Demos (1/2011) – Selections from these and others: • Deliverable D 02 - Phase 1 Final Report (1/2011) with: – Concepts (above aspects, patterns, Sys. ML basis, . . . ) – Examples (per previous slides) – Proposed next steps for Do. D applications 14

Activity 2 a Progress Leveraging existing capabilities/examples status as of 2010 -12 -13 (with completed examples listed) 15

Activity 3 a Progress Extending capabilities/examples and creating new ones status as of 2010 -12 -13 (with completed examples listed) 16

Selected Examples: Fire. Sat / NGDMC Sources: INCOSE SSWG and Inter. CAX LLC; Georgia Tech ASE 6006 18

Requirements Verification via Fire. Sat Sys. ML model “DNA signature” auto-generated from Sys. ML parametrics model Model source: Dirk. Zwemer@Inter. CAX. com 19

Selected Examples: Linkage. Systems verification pattern: unit test (two Sys. ML diagrams of to visualize same content) (A) system design being verified (T) seven (7) verification test probes wired onto system design for automated verification 20

System Design Verification Suite: Linkage. Systems “DNA signature” auto-generated from Sys. ML parametrics model (A) system design - config 2 verification pattern: multi-unit test (rolling up above unit test applied to two designs) (T) seven (7) verification test probes wired onto each system design for automated verification (A) system design - config 1 Copyright © Georgia Tech and Inter. CAX. All Rights Reserved. Sys. ML and MBSE: A Quick-Start Course 21

Selected Examples: Home Heating System Wrapped Matlab/Simulink Model – Verification Pattern verification pattern: unit test (two Sys. ML diagrams of to visualize same content) (A) Sys. ML-based system model (T) six (6) verification test probes wired onto system design for automated verification Sys. ML-based V&V added around original models by Inter. CAX LLC and Math. Works. 22

Simulation Verification Test: Home Heating System “DNA signature” auto-generated from Sys. ML parametrics model (T) six (6) test probes wired onto system design for automated verification pattern (A) system design (as wrapped Simulink model) 23

Selected Examples: Home Heating System Wrapped Matlab/Simulink Model – Sys. ML Structure (A) Sys. ML-based system model (C) Simulink model (B) Sys. ML-based wrapper (with automated interface via Para. Magic) Based on original models by Inter. CAX LLC and Math. Works. 24

Selected Examples: Wrapping Solver Models Sys. ML-wrapped system dynamics models (home heating system in Matlab/Simulink) Sys. ML-wrapped FEA models (linkage systems in Ansys) 25

System M&S Examples in STK Based on original models by AGI. 26

Summary (per SERC impact questions) • Who cares? – All M&S and VV&A stakeholders (given benefits below) • If you're successful, what difference will it make? – Our approach provides Enabling Capabilities (table rows below), which produces Primary Impacts enterprise MOEs (table columns) (measures of effectiveness) – Ex. Related earlier studies achieved 75% reduction methods/tools MOPs (measures of performance) in M&S time and enabled increased analysis intensity – We are endeavoring to demo basis for similar benefits in this SERC effort (with quantification targeted for future phases) 3/19/2018 28 28

Additional Background Material

Background • Lab/Center History @ Georgia Tech – Engineering Information Systems Lab (1996 -2006), etc. – Modeling & Simulation Lab (2006 -Present) • Director: R Peak www. msl. gatech. edu – Product & Systems Lifecycle Management Center (2005 -Present) • Director: L Mc. Ginnis Associate Directors: C Paredis and R Peak • Being renamed: Model-Based Systems Engineering (MBSE) Center • Specializations – Knowledge representations for engineering (languages, algorithms, . . . ) – Modeling & simulation interoperability – Model-based systems engineering / X (MBSE/MBX) • Sample Accomplishments – – 3/19/2018 Composable objects (became basis for Sys. ML parametrics) MRA/MIM patterns for modeling & simulation Commercialization via spin-off company: Inter. CAX LLC Contributions to related standards (Sys. ML, ISO 10303, . . . ) and organizations (INCOSE, OMG, . . . ) 30

Biosketch www. omg. org/ocsmp Russell Peak, Ph. D is a Senior Researcher at the Georgia Institute of Technology where he serves as Director of the Modeling & Simulation Lab (www. msl. gatech. edu) and Associate Director of the Product & Systems Lifecycle Management (PSLM) Center (www. pslm. gatech. edu). He is also the CTO at Inter. CAX LLC (www. Inter. CAX. com)—a spin-off company that has commercialized his work from Georgia Tech. Dr. Peak specializes in knowledge-based methods for modeling & simulation, standards-based product lifecycle management (PLM) frameworks, and knowledge representations that enable complex system interoperability. Dr. Peak originated the multi-representation architecture (MRA)—a collection of patterns for CAD-CAE interoperability—and composable objects (COBs)—a non-causal objectoriented knowledge representation. This work provided a conceptual foundation for executable parametrics in Sys. ML and for related technology commercialized by Inter. CAX in the Georgia Tech Venture. Lab program. After six years in industry (Bell Labs and Hitachi), he joined the research faculty at Georgia Tech. Since 1997 he has been principal investigator on 30+ projects with sponsors including Boeing, IBM, JPL, Lockheed, NASA, Rockwell Collins, Sandia, Shinko (Japan), TRW Automotive, US Do. C (NIST) and Do. D. He has authored over 80 publications (including several Best Paper awards), holds several patents, is an active member in ASME and INCOSE, and represents Georgia Tech on the OMG Sys. ML task force, and is a Content Developer for the OMG Certified Systems Modeling Professional (OCSMP) program. As of September 2010 he has conducted numerous Sys. ML short courses for 265+ professionals (www. pslm. gatech. edu/courses). Dr. Peak leads the INCOSE MBSE Challenge Team (www. pslm. gatech. edu/projects/incose-mbse-msi) for Modeling & Simulation Interoperability with applications to mechatronics (including mobile robotics testbeds) as a representative complex systems domain. Contact: Russell. Peak@gatech. edu Copyright © Georgia Tech and Inter. CAX. All Rights Reserved. Sys. ML and MBSE: A Quick-Start Course 31

X-Analysis Integration Techniques (c. 1993 -2004) for Modeling & Simulation Interoperability http: //eislab. gatech. edu/research/ a. Multi-Representation Architecture (MRA) b. Explicit Design-Analysis Associativity c. Analysis Module Creation Methodology © 1993 -2006 GTRC Engineering Information Systems Lab eislab. gatech. edu 32

Commercializing GIT Xai. Tools™ Technology for Executing Sys. ML Parametrics www. Inter. CAX. com Vendor Atego Sys. ML Tool Product by Inter. CAX LLC Yes Studio Prototype by GIT Para. Solver™ (2010 -2 H release) (formerly Artisan) Embedded. Plus E+ Sys. ML / RSA Yes No Magic. Draw Yes Para. Magic® (Jul 21, 2008 release) Telelogic/IBM Rhapsody — Melody™ (2010 -1 Q release) Sparx Systems Enterprise Architect n/a XMI import/export Others Yes [1] Full disclosure: Inter. CAX LLC is a spin-off company originally created to commercialize technology from RS Peak’s GIT group. GIT has licensed technology to Inter. CAX and has an equity stake in the company. RS Peak is one of several business partners in Inter. CAX. Commercialization of the Sys. ML/composable object aspects has been fostered by the GIT Venture. Lab incubator program (www. venturelab. gatech. edu) via an Inter. CAX Venture. Lab project initiated October 2007. Copyright © Georgia Tech and Inter. CAX. All Rights Reserved. Sys. ML and MBSE: A Quick-Start Course 33

Curriculum History & Formats Offered u Statistics as of Sept 2010 — www. pslm. gatech. edu/courses Full-semester Georgia Tech academic courses – ISYE / ME 8813 & 4803: Since Fall 2007 (~95 students total) u Industry short courses – Collaborative development & delivery with Inter. CAX LLC – Multiple [offerings, ~students] and formats since Aug 2008 » Sys. ML 101 [14, ~260]; Sys. ML 102 (hands-on) [12, ~205] – Modes: » Onsite at industry/government locations » Open enrollment via Georgia Tech (Atlanta, DC, Orlando, Vegas, . . . ) » Web-based “live” since Apr 2010 – Coming soon: 201/202, 301/302 (int/adv concepts, OCSMP prep, . . . ) u Georgia Tech Professional Masters academic courses – Professional Masters in Applied Systems Engineering www. pmase. gatech. edu – ASE 6005 Sys. ML-based MBSE course - Summer 2010 – ASE 6006 SE Lab (Sys. ML-based system design project) - Fall 2010 Copyright © Georgia Tech and Inter. CAX. All Rights Reserved. Sys. ML and MBSE: A Quick-Start Course 35

Mobile Robot Exercise from myro import * initialize("com 29") senses() Executable Sys. ML Activity Model [after live update] beep(1, 440) Resulting python script Copyright © Georgia Tech and Inter. CAX. All Rights Reserved. Sys. ML and MBSE: A Quick-Start Course forward(1, 1) turn. Right(1, . 4) forward(1, 1) beep(1, 440) turn. Right(1, . 4) forward(1, 1) stop() 39

Decision Nodes / Guard Conditions and Merge Nodes decision node guard condition (with sensor reading) merge node* Copyright © Georgia Tech and Inter. CAX. All Rights Reserved. Sys. ML and MBSE: A Quick-Start Course 40