Dov Dori Massachusetts Institute of Technology - PowerPoint Presentation

Slide1

Dov DoriMassachusetts Institute of TechnologyTechnion, Israel Institute of TechnologyOct. 14, 2015

Knowledge Based Engineering withOPM – the New ISO 19450 standard

Slide2

What is required of an agile conceptual modeling language?It must be:Simple

(quickly learnable & understandable) yet expressive (capable of saying a lot with little)Intuitive (for humans) yet formal (for machines)Seemingly contradicting requirements!How can these be reconciled?

Slide3

A conceptual

modeling language that issimple yet expressive, and intuitive

yet

formal

Let the search begin!

Slide4

Preamble: OCCUM’s RAZOR

"If you have two equally likely solutions to a problem, choose the simplest.""Entities should not be multiplied unnecessarily." "Entia non sunt multiplicanda praeter necessitatem"14th Century logician and Franciscan priest William of Ockham In an extended version:OCCUM’s RAZOR is an important

guiding

principle

of OPM

Slide5

Universal Ontology

Ontology:

a set of concepts for describing a domain (industry, banking, military, botany, healthcare…) and systems within it.

Universal Ontology:

a

domain-independent

set

of concepts for describing

systems in the universe, both natural

and

man-made.

Slide6

Fundamental question 1:

What is needed to describe the universe? Answer: Describing the universe requires things and relations among them.

10/14/2015

Slide7

Question 2:

What is a thing or what can it do? Answer:

A things can either

exist

or

happen

.

10/14/2015

Any thing either

exists

or

happens

– nothing else!

Slide8

Question 3:

What are the things that exist in the world? Answer: Objects exist.They are

static

–

time independent.

10/14/2015

Slide9

Question 4:

What are the things that happen in the world? Answer: Processes happen.They are dynamic – time dependent.

10/14/2015

Slide10

Question 5:

How do objects and processes relate?Answer: Processes happen to objects.While happening,

processes

transform

objects.

10/14/2015

Slide11

Question 6:

What does a process do when it happens to an object?Answer: The process transforms theObject.

10/14/2015

Slide12

OPM Things

: Objects and Processes

Object:

A

thing that

exists or might exist

physically or

informatically.

Process:

A

thing that transforms one or more

objects.

12

Slide13

OPM’s only two building blocks:

1.

Stateful

Object

2.

Process

All the other elements are relations between things, expressed graphically as links.

13

Slide14

processes transform

objects.Question 7: What does transforming mean?Transforming means creating an

object

or

destroying

an

object

or

affecting

an

object.

10/14/2015

Slide15

15

Transforming an object by a process can be done in three ways

(1)

Process consumes the object

Slide16

16

(2)

Process creates the

object

Creation

Consumption

Slide17

processes affect objects.

Question 8: What does affecting mean?

10/14/2015

A

process

affects

an

object

by

changing its

state.

Hence,

objects

must be

stateful

–

they must have

states.

Slide18

18

(3)

Process affects object by changing the object’s state

:

The third and last kind of object transformation:

Slide19

The three transformation kinds

Consumption:Creation:State Change:

OPM uses a single type of diagram –

Object-Process

Diagram (OPD)

Graphic edit operations are translated

on the fly to natural language –

Object-Process Language (OPL)

Catering to dual

channel

processing

Slide20

20

The

graphics-text

equivalence OPM

principle

Any

model fact expressed graphically in an OPD is also expressed textually in the corresponding OPL paragraph.

Caters to the dual channel cognitive assumption (Mayer, 2010)

Slide21

Physical vs. Informatical Things

OPCAT – downloadable free from http://esml.iem.technion.ac.il/

Slide22

The Object-Process Theorem

Stateful objects, processes, and relations among them constitute a necessary and sufficient universal ontology.

10/14/2015

Slide23

Question 9: What are

the two major aspects of any system?Structure – the static aspect: what the system is made of. Time-independentBehavior – the dynamic aspect: how the system changes over time.Time-dependent

10/14/2015

Slide24

Question 10: What third aspect is specific to man-made systems?

Function – the utilitarian, subjective aspect: Why is the system built? For whom is the system built?Who benefits from operating the system?VALUE IS BENEFIT AT COST Benefit comes from function (processes)Cost comes from form (hardware, stuff, objects)

10/14/2015

Slide25

Question 11: What is the Value which the beneficiary seeks?

VALUE IS BENEFIT AT COSTVALUE = BENEFIT – COST Benefit

comes from function (processes).

We want to

maximize

it

Cost

comes from form (hardware, stuff, objects).

We want to

minimize

it

10/14/2015

Slide26

Question 12: What is System Architecting & Architecture?

System architecting is mapping form to function to maximize value to the system’s beneficiarySystem architecture is the combination of structure (objects) and behavior (processes) that maximize the value to the beneficiary

10/14/2015

Slide27

Two Complementary Proofs:

1. Theoretical, based on logic2. Empirical, based on examples

10/14/2015

Slide28

Theoretical Proof Part 1

- Necessity

10/14/2015

Stateful objects

and

processes

are

necessary

to specify the two system aspects

:

Specifying the

structural

, static system aspect

requires

stateful

objects

and relations among them

.

Specifying the

procedural

, dynamic system aspect

mandates using

processes

and relations between them and the

objects

they transform.

Slide29

Stateful objects and processes

are sufficient to specify any thing in any system:Anything that exists or might exist can be specified in terms of stateful objects and relations among them.

10/14/2015

Anything that

happens

or might happen to any stateful object can be specified in terms of

processes

and relations between them and the object they transform.

Q.E.D.

Theoretical Proof

Part

2

-

Sufficiency

Slide30

Empirical Proof of the Object-Process Theorem

Stateful objects, processes, and relations among them constitute a necessary and sufficient universal ontology.If the ontology is universal, it can model systems in any domain.The empirical proof: Providing evidence of successful models from various, unrelated domains.

10/14/2015

Slide31

Empirical Proof from Science: Molecular biology

10/14/2015

Slide32

10/14/2015

32

“Beyond

the scientific value of these specific findings, this work demonstrates the value of the conceptual model as an in silico vehicle for hypotheses generation and testing, which can reinforce, and often even replace, risky, costlier wet lab experiments

.”

Slide33

Tripped Pumps Cause too high Pressure

Nuclear reactor failure:

The Three Mile Island Accident

Slide34

10/14/2015

Offshore Oil Well Drilling

Slide35

10/14/2015

Airport Operations: Outgoing Passenger

Slide36

10/14/2015

Iron

Dome – an Israeli ballistic missile defense system

Yaniv Mordecai and Dov Dori, Evolving System Modeling: Facilitating Agile System Development with Object-Process Methodology.

SysCon 2015

,

9

th

Annual IEEE International Systems Conference

, Vancouver, Canada, April 13-16 2015.

To be presented

Slide37

Sample of engineering domains in which OPM has been used

Complex, Interconnected, Large-Scale Socio-Technical Systems. Systems Engineering 14(3), 2011.Networking Mobile Devices and Computers in an Intelligent Home. International Journal of Smart Home 3(4), pp. 15-22, October, 2009.

Multi-Agent Systems

.

IEEE Transactions on Systems, Man, and Cybernetics – Part C: Applications and Reviews

, 40 (2) pp. 227-241, 2010

.

Semantic Web Services Matching and Composition. Web Semantics: Science, Services and Agents on the World Wide Web

. 9, pp. 16-28, 2011.Project-Product Lifecycle

Management

.

Systems Engineering

, 16 (4), pp. 413-426, 2013

.

Model-Based

Risk-Oriented

Robust

Systems Design

.

International Journal of Strategic Engineering Asset Management

, 1(4), pp. 331-354, 2013

.

Medical Robotics

and Miscommunication Scenarios.

An Object-Process Methodology Conceptual Model.

Artificial Intelligence in Medicine

, 62(3) pp. 153-163, 2014

.

Modeling Exceptions in

Biomedical Informatics

.

Journal of Biomedical Informatics

42

(4), pp. 736-747, 2009.

Slide38

Complexity Management with OPMSystems are inherently complex.To alleviate this complexity, in OPM, it is managed by detail decomposition through three

refinement-abstraction:In-zooming – Out-zoomingUnfolding – FoldingState expression – suppression. 38

Slide39

39

In-zooming – Out-zooming ExampleProcess Performance Controlling- a metamodel from ISO 19450

All the OPDs, at any detail level, are self-similar.

They contain only stateful objects, processes, and relations.

Slide40

OPM Complexity Management BenefitsThere is no limit on the level of complexity of the system being modeled:One can specify system structure and behavior at any level

of detail by recursively in-zooming. Catering to the cognitive limited capacity:Each diagram is not overly complicated.All the diagrams are “aware” of each other:All OPDs are partial views of the same system.Any change in one diagram is propagated to all the other relevant ones. 40

Slide41

Whirlpool New Gen Dishwasher: Model-Based Design Outline

Model requirement

with customer (or Marketing as proxy)

This is the

problem domain model

Specify alternative selection criteria based on value as benefit as cost

This will enable decision

making once we have

alternative concepts

Slide42

Whirlpool New Gen Dishwasher: Model-Based Design Outline

Formulate 2-3 alternative concepts

This is the solution model which evolves from and extends the problem model

Each

concept

shall be based

on a different underlying idea/physical principle

Model each concept with OPM to level 1 or 2 (SD is level 0)

Slide43

Whirlpool New Gen Dishwasher: Model-Based Design Outline

Evaluate the concept models

based on the predefined criteria

Select the concept model with the best value

Continue modeling the selected concept to more refined subsystem & component levels

Where applicable, repeat developing alternatives for subsystems (steps 4-7).

Slide44

Whirlpool New Gen Dishwasher: Model-Based Design Outline

Involve customer or marketing throughout the modeling process

Stop when the conceptual model is sufficiently detailed and prescriptive for transition to hardware/software/ GUI/exterior/packaging design

Design

the

PPLM

model with built-in

tests for meeting

all the requirements (not in this workshop)

Slide45

Requirements Modeling –

The Problem Domain OPM model

What is the

function

of this system?

Describe in three words, the last being a verb ending with ing (gerund)

This will be our starting point of the requirements OPM model

Slide46

The Function:

HouseholdDish Caring

Slide47

Who is the Beneficiary?

Who benefits from owning and operating the system?

Slide48

What is the Operand (the object that the function transforms)?

Who benefits from owning and operating the system?What is the attribute of the operand whose change generates value to the beneficiary?

Slide49

What are the system’s input and output?

Slide50

What is the name of the new system we are developing?

Slide51

The next detail level: Zooming into the

System’s Function

Slide52

The Auto-Generated OPL Text:

A self-documenting feature

Slide53

OSRVT: Video Moving Target Indication Capability

Presented to SRI InternationalNovember 10, 2014

Slide54

OSRVT: One System Remote Video Terminal System54

MTITextron’s One System Remote Video Terminal, a laptop soldiers on the ground can use to see a drone’s video and control its sensors.

Slide55

ObjectiveAdding Video Moving Target Indication* (VMTI) capability for OSRVT using platform data stream (e.g., MPEG2 TS)*A computational process of locating a moving object (or several ones) in a video frame. No ID reported

.Note: Introduction of such a capability will have little or no impact on other OSRVT operations.55

Slide56

56

VMTI CSCI

HMI CSCI

(Front

End)

(Back End)

OSRVT VMTI System

Video Moving Target Indication

OSRVT: One System Remote Video Terminal System

Slide57

VMTI Module57

Image QualityMTI

Input

Imagery

Metadata

Screened

Imagery

VMTI Imagery

MTI

ROI’s

(per frame)

F2F Alignment

Input Imagery

Con

t

r

o

l

l

e

r

MTI

Params

F2F Alignment

STAB

Screened

Imagery

Screened

Imagery

F2F Alignment

O

v

e

r

l

a

y

D

e

t

e

c

t

Overlay

Mask

Enable/Disable

OSRVT

OSRVT

(from)

(to)

Generates MTI

params

from metadata

Screens input frames for image defects

Aligns consecutive frames for stabilized stream

Direct method

Feature based

Detects image overlays, generates mask

Detects moving targets outside of overlay mask in screened, stabilized imagery

Synchronized

Options

Slide58

Moving Target Indication58

Video frames (stabilized) with moving targetsMoving Target Indication CoreDetection

Verification

Blob

Extraction

Laplacian

Pyramid

Generation

Change/

Foreground

Detection

Blob

Filtering

+

Overlay mask imagery

Video frames

(original)

with

moving

target indications

(VMTI imagery)

+

MTI parameters

Generates multi-resolution features (

Laplacian

)

Detects pixel-based

spatio

-temporal changes,

relative to stabilized “

background,” due to “foreground” features from a particular pyramid level

(as specified by MTI parameters)

Extracts “blobs” or connected foreground

“change”

pixels

Filters “blobs” fulfilling motion consistency (temporal) check

Slide59

MTI Core Details59

Compute spatiotemporal image gradientsTemporalFiltering for Consistent PixelsCompute Optical Flow

F

n

RF

n

NF

n

MF

n

FF

n

F

n

: Histogram equalized frame w/ stab

params

RF

n

:

Laplacian

of

F

n

+ Stabilized ref frame

NF

n

:

F

n

+

Binarized

F

n

of change pixels

MF

n

:

F

n

+

Binarized

F

n

of consistent change pixels

Connected

Component

Labeling

BF

n

Laplacian

Pyramid

Generation

Normal Flow Based Change Detection

Compute

Normal Flow

values

Threshold

Normal Flow

values

Blob Extraction

Extract

pixel

groups

using labels

Optical Flow Based Blob Filtering

Reverse flow

warp Insp Blobs,

Check Temporal

Consistency

P

F

n

Parallax Detection (optional)

Visual Odometry

(Camera R, T)

From Optical Flow

De-rotate

Optical Flow

Field

Compute Epipole

And Enforce

Epipolar Constraint

on Blob Pixels

Update Blobs

After Removing

Pixels Failing

Epipolar Constraint

BF

n

:

F

n

+ Blobs of change pixels

FF

n

:

F

n

+ consistent

blobs + Optical Flow Field

P

F

n

:

F

n

+

consistent blobs (following Parallax Detection)

Adaptive Background

Modeling

Foreground/Background

Segmentation

Foreground

Detection (optional)

Video frames (stabilized)

with

moving

targets

+

Overlay mask imagery

+

MTI parameters

Video frames

(original)

with

moving

target indications

(VMTI imagery)

Check

Motion

Significance

Slide60

Moving

Target Indicator: What is the Function of this system?

Describe in three words, the last being a verb ending with ing (gerund)

This will be our starting point of the OPM model

Slide61

The Function:

Moving Target Indicating

Slide62

Who is the Beneficiary?

Who benefits from operating the system?

Slide63

What attribute of War Fighter changes value by operating the

system, such that benefit is created?

Slide64

What are the system’s input and output?

Slide65

What is the name of the system we are developing?

Slide66

The next detail level: Zooming into the

Moving Target Indicating Function

Slide67

The Auto-Generated OPL Text:

A self-documenting feature

Slide68

The Next Level of Detail will be based on this:

Slide69

Summary: OPM Aspect Unification

The three system aspects:Function (why the system is built), Structure (static aspect: what is the system made of), and Behavior (dynamic aspect: how the system changes over time)Are expressed bi-modally, in graphics and equivalent textIn a single model

Slide70

Agile OPM-MBSE Highlights Model the requirements

together with the customerUse this model as a basis for concept generation and their evaluation and selection of best oneAchieve shared understanding and agreement of multidisciplinary engineering teamCommunicate the solution model with the customer Evolve and use the model across all the system lifecycle: detailed design, integration, testing, deployment, maintenance, retirement…

Slide71

OPM Resources:

Book: Object-Process Methodology - A Holistic Systems Paradigm, Springer Verlag, Berlin, Heidelberg, New York, 2002.Upcoming book (2015)

Model-Based Systems Engineering with OPM and SysML

,

Springer,

New

York.

Standard ISO/PAS 19450

OPMWebsite: Enterprise Systems Modeling Laboratory

contains

journal & conference papers,

free OPCAT software,

presentations,

projects, and more.

Slide72

Appendix: SysML and OPM – a brief comparison Feature

SysMLOPMTheoretical foundationUML; Object-Oriented paradigmMinimal universal ontology;Object-Process TheoremStandard documentation number of pages1670 (700 + 700 + 270)130(100 + 30)Standardization body

OMG (2006)

ISO (2014)

Number of diagram kinds

9

1

Graphic modality

yes

yes

Textual modality

no

yes

Physical-informatical distinction

no

yes

Systemic-environmental distinction

no

yes

Slide73

Questions and

(hopefully) Answers

Join the growing OPM community

Here

!

https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=OPM&A=1

Contact: Dov Dori –

dori@mit.edu