1

Cyber-Physical System Ensembles:

Unlocking opportunities when machines collaborate

Pieter J. Mosterman

Chief Research Scientist, Director

MathWorks

Justyna Zander

Adjunct Professor

School of Computer Science, McGill University

MathWorks Research Fellow

Worcester Polytechnic Institute

Axes of Classification

Configure online Collaborate Infrastructure

Needs, challenges, technological solutions, and the potential impact are classified along three axes: (i) dynamic online

configuration of these systems of systems, (ii) forms of collaboration by the machines, and (iii) infrastructural support for the

collaboration of dynamically configuring system ensembles

Information

Physics

Electronics

Network

Physics

Information

Electronics

Network

Confidently design systems as part of a reliable system

ensemble

Exploit exogeneous functionality for efficient, economical, and

resilient operation

Configure online

Configure online

4

Linearization, implementation model generation, etc.

Generation of models with necessary detail based on property

selection

Proper models in design ►

Confidently design systems as part of a reliable system

ensemble

Exploit exogeneous functionality for efficient, economical, and

resilient operation

Configure online

Virtual system integration

Need Technology

Impact

Challenge

5

Linearization, implementation model generation, etc.

Generation of models with necessary detail based on property

selection

Model Building

Automation SystemCounterexample

guided abstraction

refinement

Proper models in design ►

Confidently design systems as part of a reliable system

ensemble

Exploit exogeneous functionality for efficient, economical, and

resilient operation

Virtual system integration

Configure online

6

Link for cosimulation,

simulation API, code

generation

Solver configurations for

continuous time, discrete

time, discrete event

Connecting, combining, and

integrating models represented in

different formalisms

Efficient simulation models to be

used across dynamic and execution

semantics

System-level design and analysis by using models ►

Confidently design systems as part of a reliable system

ensemble

Exploit exogeneous functionality for efficient, economical, and

resilient operation

Virtual system integration

Need Technology

Impact

Challenge

Configure online

7

Link for cosimulation,

simulation API, code

generation

Solver configurations for

continuous time, discrete

time, discrete event

Connecting, combining, and

integrating models represented in

different formalisms

Efficient simulation models to be

used across dynamic and execution

semantics

Hybrid dynamic

systems

Multiparadigm modeling

System-level design and analysis by using models ►

Confidently design systems as part of a reliable system

ensemble

Exploit exogeneous functionality for efficient, economical, and

resilient operation

Virtual system integration

Configure online

8

Data streaming, target connectivity support, standardized

communication protocols (TCP, UDP), real-time simulation

Open tool platforms with trusted interfaces for communication

across synchronized and coordinated models, software, and

hardware devices

Connectivity among models, software, and hardware

Confidently design systems as part of a reliable system

ensemble

Exploit exogeneous functionality for efficient, economical, and

resilient operation

Virtual system integration

Need Technology

Impact

Challenge

Configure online

9

Data streaming, target connectivity support, standardized

communication protocols (TCP, UDP), real-time simulation

Open tool platforms with trusted interfaces for communication

across synchronized and coordinated models, software, and

hardware devices

Connectivity among models, software, and hardware

Confidently design systems as part of a reliable system

ensemble

Exploit exogeneous functionality for efficient, economical, and

resilient operation

Virtual system integration

Real-time simulation

Configure online

10

Handling of ensemble

(in)consistency with

sufficient runtime

fidelity

Traceability (direct and

across

transformations)

Runtime variants,

middleware service

description

specification (.srv)

Runtime curve fitting

and design

optimization

Introspection of the

system state,

configuration, and

available services

Online model

calibration

Reasoning and planning adaptation of an ensemble of systems ►

Confidently design systems as part of a reliable system

ensemble

Exploit exogeneous functionality for efficient, economical, and

resilient operation

Runtime system adaptation

Need

Impact

Technology

tem

Challenge

Configure online

11

Handling of ensemble

(in)consistency with

sufficient runtime

fidelity

Traceability (direct and

across

transformations)

Runtime variants,

middleware service

description

specification (.srv)

Runtime curve fitting

and design

optimization

Introspection of the

system state,

configuration, and

available services

Online model

calibration

Reasoning and planning adaptation of an ensemble of systems ►

Confidently design systems as part of a reliable system

ensemble

Exploit exogeneous functionality for efficient, economical, and

resilient operation

Runtime system adaptation

Models @ runtime Automated model

calibration

Configure online

12

Regression modeling, model selection (artificial neural network,

support vector machine, rational model)

Environment models to enable surrogate interactions

Testing with functionality on deployed systems

Confidently design systems as part of a reliable system

ensemble

Exploit exogeneous functionality for efficient, economical, and

resilient operation

Runtime system adaptation

TechnologyChallengeNeed

Impact

Configure online

13

Regression modeling, model selection (artificial neural network,

support vector machine, rational model)

Environment models to enable surrogate interactions

Testing with functionality on deployed systems

Confidently design systems as part of a reliable system

ensemble

Exploit exogeneous functionality for efficient, economical, and

resilient operation

Runtime system adaptation

Testing with surrogate

models

Configure online

14

Information

Physics

Electronics

Network

Physics

Information

Electronics

Network

Collaborate

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Collaborate

15

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Emerging behavior design

Planning and synthesis

of distributed control

functionality on

concurrent resources

Concurrency and

platform modeling,

functionality

decomposition, service

composition

Event-driven control,

discrete event

modeling and analysis,

uncertainty modeling

Concurrency

semantics, property

proving with

performance models

Analysis methods

across loosely coupled

architectures

Accessible formal

methods that apply to

collaborative problems

Reasoning and planning adaptation of an ensemble of systems

Need

Impact

Technology

tem

Challenge

Collaborate

16

Planning and synthesis

of distributed control

functionality on

concurrent resources

Concurrency and

platform modeling,

functionality

decomposition, service

composition

Event-driven control,

discrete event

modeling and analysis,

uncertainty modeling

Concurrency

semantics, property

proving with

performance models

Analysis methods

across loosely coupled

architectures

Accessible formal

methods that apply to

collaborative problems

Service orchestrationService oriented sensor

programming

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Reasoning and planning adaptation of an ensemble of systemsEmerging behavior design

Collaborate

17

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Data sharing

Communication modeling, double buffering analysis, timing

properties of software, clock recovery

Synchronization of data from incongruent

sources

Multirate architectures ►

TechnologyChallengeNeed

Impact

Collaborate

18

Communication modeling, double buffering analysis, timing

properties of software, clock recovery

Synchronization of data from incongruent

sources

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Data sharing Multirate architectures ►

Technology challenges in

CPS

Collaborate

19

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Data sharing

Version based model import/export, metamodel generation, model

concepts sharing and comparing

Information represented as high-level models with well-defined

metamodels and ontologies

Extracting and deriving specific value from general information

TechnologyChallengeNeed

Impact

atio

Collaborate

20

Version based model import/export, metamodel generation, model

concepts sharing and comparing

Information represented as high-level models with well-defined

metamodels and ontologies

Megamodeling and

metamodeling

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Data sharing Extracting and deriving specific value from general information

Collaborate

21

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Performance

characterization via

performance models

and measures

Critical path analysis,

code performance

report and advisor

Property based model

slicing, behavioral

analysis, functionality

mining

Adaptive filtering,

distortion modeling,

groundtruthing

(baselining)

Generation of models

for a task by property

identification and

model behavior

selection

Online calibration

based on objective and

performance criteria

Multi-use functionality post-deploymentFunctionality sharing ►

Collaborate

22

Performance

characterization via

performance models

and measures

Critical path analysis,

code performance

report and advisor

Property based model

slicing, behavioral

analysis, functionality

mining

Adaptive filtering,

distortion modeling,

groundtruthing

(baselining)

Generation of models

for a task by property

identification and

model behavior

selection

Online calibration

based on objective and

performance criteria

Multi-use functionality post-deployment ►

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Functionality sharing

Data distribution

service (DDS)

Online calibrationRequirements mining

Collaborate

23

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Property and assumption based model slicing, trace to source and

destination, assumptions in functionality to behavior mapping

Assumption formalization and dependency effect analysis

Feature interactionFunctionality sharing

Collaborate

24

Property and assumption based model slicing, trace to source and

destination, assumptions in functionality to behavior mapping

Assumption formalization and dependency effect analysis

Feature interaction

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Functionality sharing

Multi-rate double

buffering

Collaborate

25

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Coverage based automatic test generation, variants-based

testing, closed-loop testing

Model-based test generation from requirements while preserving

the context of dynamic configuration

Systematic test suite generation and automated test evaluationCollaborative functionality testing ►

Collaborate

26

Coverage based automatic test generation, variants-based

testing, closed-loop testing

Model-based test generation from requirements while preserving

the context of dynamic configuration

Systematic test suite generation and automated test evaluation ►

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Collaborative functionality testing

Testing of dynamic

variability

Collaborate

27

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Collaborative functionality testing

System state restoration, stateless

services, test fixture generation

Time partition testing, functionality

extraction

Setting of initial conditions and

injecting fault data

Temporal and spatial partitioning to

isolate functionality for a specific

system architecture under

investigation

Reproducible test results under minimum uncertainty

Collaborate

28

System state restoration, stateless

services, test fixture generation

Time partition testing, functionality

extraction

Setting of initial conditions and

injecting fault data

Temporal and spatial partitioning to

isolate functionality for a specific

system architecture under

investigation

Reproducible test results under minimum uncertainty

Systematically design

systems that are part of a

system ensemble to optimally

realize desired ensemble

behavior

Effectively exploit distributed

information resources for

exclusive system features

Create novel system features

post deployment

Assure the collaboration

quality on shared resources.

Identify and automatically

mitigate root causes of failure

in a distributed environment.

Collaborative functionality testing

Service oriented

architecture testing

Collaborate

29

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Information

Physics

Electronics

Network

Physics

Information

Electronics

NetworkInfr

astr

uctu

re

Infrastructure

30

Relations (across abstractions,

formalisms, transformations)

service API, change notification API

Protected models (obfuscated,

encrypted), trusted compiler

Traceability across semantic and

technology adaptation, and

intellectual property protection

Information extraction from

obfuscated intellectual property

Tool coupling among disparate organizations ►

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Design artifact sharing

Infrastructure

31

Relations (across abstractions,

formalisms, transformations)

service API, change notification API

Protected models (obfuscated,

encrypted), trusted compiler

Traceability across semantic and

technology adaptation, and

intellectual property protection

Information extraction from

obfuscated intellectual property

Tool coupling among disparate organizations ►Design artifact sharing

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Open services for

lifecycle collaboration

Infrastructure

32

Model generation, pattern (control)

extraction, XML interexchange

Modeling numerical mathematics

(integration, root finding) as

dynamic system

Configurable view projections that

are tool specific

Consistent semantics across tools

by modeling the execution engines

Support manifold views and tools in design

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Design artifact sharing

Infrastructure

33

Model generation, pattern (control)

extraction, XML interexchange

Modeling numerical mathematics

(integration, root finding) as

dynamic system

Configurable view projections that

are tool specific

Consistent semantics across tools

by modeling the execution engines

Support manifold views and tools in design

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Graph transformationsSingle underlying model

Design artifact sharing

Infrastructure

34

Communication protocol (building block) modeling, performance

modeling across target hardware

Real-time services of graded quality with a low footprint and a

configurable protocol stack that includes time and location

information

Physically aware configurable protocol stack that is IP compatible►Wireless communication

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Infrastructure

35

Communication protocol (building block) modeling, performance

modeling across target hardware

Real-time services of graded quality with a low footprint and a

configurable protocol stack that includes time and location

information

Physically aware configurable protocol stack that is IP compatible►Wireless communication

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

IEEE 802.15.4e low cost

communication

Infrastructure

36

Wireless communication

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Physical layer (RF) modeling,

antenna modeling

Scheduler configuration, dynamic

scheduling with guarantees, mixed

synchronous and asynchronous

behavior

Physical layer based timing and

synchronization architectures

Scheduling of periodic and

aperiodic events with reliable

execution times

Precise timing and synchronization in a distributed environment

Infrastructure

37

Physical layer (RF) modeling,

antenna modeling

Scheduler configuration, dynamic

scheduling with guarantees, mixed

synchronous and asynchronous

behavior

Physical layer based timing and

synchronization architectures

Scheduling of periodic and

aperiodic events with reliable

execution times

Precise timing and synchronization in a distributed environmentWireless communication

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

IEEE 1588 precise timing Distributed real-time

systems task scheduling

Processor and

network scheduling

Infrastructure

38

Virtual machine (LLVM, JVM, Docker) with real-time capabilities,

serialized intermediate representation of functionality

Standardized and configurable real-time execution stack

Flexible and transferable embedded functionality dispatch ►

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Hardware resource sharing

Infrastructure

39

Virtual machine (LLVM, JVM, Docker) with real-time capabilities,

serialized intermediate representation of functionality

Standardized and configurable real-time execution stack

Flexible and transferable embedded functionality dispatch ►

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Hardware resource sharing

Real-time virtualizationOpen Services Gateway

Initiative (OSGi)

Infrastructure

40

Combine analytic and experimental target profiling (processor,

hardware, FPGA -in-the-loop), interpolation estimates from

historical data

Platform-based modeling of execution behavior functionality

Performance characterization from abstract functionality ►

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Hardware resource sharing

Infrastructure

41

Combine analytic and experimental target profiling (processor,

hardware, FPGA -in-the-loop), interpolation estimates from

historical data

Platform-based modeling of execution behavior functionality

Performance characterization from abstract functionality ►

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Hardware resource sharing

Platform-based design

Infrastructure

42

Static analysis methods, automatic test generation, hardware

architecture behavior implementation

Characterization of computational architectures

Determination of key test cases for different implementations ►

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Hardware resource sharing

Infrastructure

43

Static analysis methods, automatic test generation, hardware

architecture behavior implementation

Characterization of computational architectures

Determination of key test cases for different implementations ►

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Hardware resource sharing

IEEE 754 floating point

denormals

Infrastructure

44

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Hardware resource sharing

Harmonic periods (integer),

synchronous (single clock, discrete

time), simultaneous, dense

(variable step, continuous)

Certification kit for mixed Safety-

Integrity Levels (SiL) of

components, matching software

with hardware

Modeling the semantics of timeDynamically mixing safety integrity

levels

Safety of heterogeneous system ensembles

Infrastructure

45

Harmonic periods (integer),

synchronous (single clock, discrete

time), simultaneous, dense

(variable step, continuous)

Certification kit for mixed Safety-

Integrity Levels (SiL) of

components, matching software

with hardware

Modeling the semantics of timeDynamically mixing safety integrity

levels

Safety of heterogeneous system ensembles

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Hardware resource sharing

Safety analysis at an

architectural level

Infrastructure

46

Service discovery response time (latency, averages, time-out),

modal service request behavior

Real-time middleware and service oriented architectures with

physical capabilities

Real-time embedded services operating in a physical environment►

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Service utilization

Infrastructure

47

Service discovery response time (latency, averages, time-out),

modal service request behavior

Real-time middleware and service oriented architectures with

physical capabilities

Real-time embedded services operating in a physical environment►

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Service utilization

Real-time discovery

services

Real-time middleware

Infrastructure

48

Type similarity checking and conversion, semantics definition

Service ontologies with taxonomies for similarity and

transformability matching

Smart services discovery ►

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Service utilization

Infrastructure

49

Type similarity checking and conversion, semantics definition

Service ontologies with taxonomies for similarity and

transformability matching

Smart services discovery ►

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Service utilization

Semantic middleware

Infrastructure

50

Reliable model and code generation

Language and ontology infrastructure to support translation and

transformation

Information sharing in a heterogeneous system ensemble

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Service utilization

Infrastructure

51

Reliable model and code generation

Language and ontology infrastructure to support translation and

transformation

Information sharing in a heterogeneous system ensemble

Conveniently, efficiently, and

consistently collaborate

between stakeholders

throughout the system life

cycle

Reliably configure flexible

system configurations for

features with varying quality

of service

Contract out system

resources and balance use of

external resources for

resiliency and runtime cost

optimization

Dynamically assemble

systems post-deployment

and purpose available

functionality to serve singular

needs

Service utilization

Semantic web

ontology language

Semantic anchoring

Infrastructure

52

Pieter J. Mosterman and Justyna Zander, “Cyber-physical

systems challenges: a needs analysis for collaborating

embedded software systems,” in Software & Systems

Modeling, Springer Berlin/Heidelberg, ISSN 1619-1366, vol.

15, nr. 1, pp. 5-16, 2016

https://dl.acm.org/citation.cfm?id=2890224 https://www.youtube.com/watch?v=oofHMaEWwP8

The Smart Emergency Response System

Using MATLAB and Simulink

Environment models to enable surrogate interactions

Open tool platforms with trusted interfaces for communication

across synchronized and coordinated models, software,

and hardware devices

Efficient simulation models to be used across dynamic and execution semantics

Online model calibration

Maintenance of consistent information and management of inconsistencies regarding the ensemble of systems with sufficient fidelity at runtime

Introspection of the system state, configuration, and services it makes available

Generation of models with necessary detail based on property selection

Connecting, combining, and integrating models represented in different formalisms

Needs Challenges Technology ImpactD

yn

am

icall

y C

on

fig

uri

ng

Cyb

er-

Ph

ysi

cal Syst

em

En

sem

ble

s

Pie

ter

J. M

ost

erm

an

an

d J

ust

yn

a Z

an

der, “

Cyb

er-

Ph

ysi

cal Syst

em

s C

hallen

ges—

A N

eed

s A

naly

sis

for

Co

llab

ora

tin

g E

mb

ed

ded

So

ftw

are

Syst

em

s,”

in S

oft

ware

&

Syst

em

s M

odelin

g, Sp

rin

ger

Berl

in/H

eid

elb

erg

, IS

SN

1619-1

366. v

ol.

15, n

r. 1

, p

p.

5-1

6, 2016

[htt

p:/

/msd

l.cs.

mcg

ill.c

a/p

eo

ple

/mo

sterm

an

/pap

ers

/so

sym

15/p

.pd

f]

Model-based test generation from requirements while preserving the context of dynamic configurations

Setting of initial conditions and injecting fault data

Temporal and spatial partitioning to isolate

functionality for a specific system architecture under

investigation

Information represented as high-level models with well-defined

metamodels and ontologies

Synchronization of data from incongruent sources

Assumption formalization and dependency effect analysis

Online calibration based on objective and performance criteria

Performance characterization via performance models and measures

Generation of models for a desired task perspective by property identification and model behavior selection

Accessible formal methods that apply to collaborative problems

Planning and synthesis of distributed control functionality on concurrent resources

Analysis methods across loosely coupled architectures

Needs Challenges Technology Impact

Pie

ter

J. M

ost

erm

an

an

d J

ust

yn

a Z

an

der, “

Cyb

er-

Ph

ysi

cal Syst

em

s C

hallen

ges—

A N

eed

s A

naly

sis

for

Co

llab

ora

tin

g E

mb

ed

ded

So

ftw

are

Syst

em

s,”

in S

oft

ware

&

Syst

em

s M

odelin

g, Sp

rin

ger

Berl

in/H

eid

elb

erg

, IS

SN

1619-1

366. v

ol.

15, n

r. 1

, p

p.

5-1

6, 2016

[htt

p:/

/msd

l.cs.

mcg

ill.c

a/p

eo

ple

/mo

sterm

an

/pap

ers

/so

sym

15/p

.pd

f]

Co

llab

ora

tin

g

Cyb

er-

Ph

ysi

cal Syst

em

En

sem

ble

s

Real-time middleware and service oriented architectures with physical capabilities

Service ontologies with taxonomies for similarity and transformability

matching

Language and ontology infrastructure to support

translation and transformation

Scheduling of periodic and aperiodic events with reliable

execution times

Physical layer based timing and synchronization architectures

Real-time services of graded quality with a low footprint and a configurable protocol stack that

includes time and location information

Characterization of computational architectures

Dynamically mixing safety integrity levels

Modeling the semantics of time

Platform-based modeling of execution behavior functionality

Standardized and configurable real-time execution stack

Traceability across semantic and technology adaptation, and intellectual property protection

Information extraction from obfuscated intellectual property

Configurable view projections that are tool specific

Consistent semantics across tools by modeling execution engines

Pie

ter

J. M

ost

erm

an

an

d J

ust

yn

a Z

an

der, “

Cyb

er-

Ph

ysi

cal Syst

em

s C

hallen

ges—

A N

eed

s A

naly

sis

for

Co

llab

ora

tin

g E

mb

ed

ded

So

ftw

are

Syst

em

s,”

in S

oft

ware

&

Syst

em

s M

odelin

g, Sp

rin

ger

Berl

in/H

eid

elb

erg

, IS

SN

1619-1

366. v

ol.

15, n

r. 1

, p

p.

5-1

6, 2016

[htt

p:/

/msd

l.cs.

mcg

ill.c

a/p

eo

ple

/mo

sterm

an

/pap

ers

/so

sym

15/p

.pd

f]Needs Challenges Technology ImpactIn

frast

ructu

refo

r

Cyb

er-

Ph

ysi

cal Syst

em

En

sem

ble

s