cerbero in a nutshell overview of consortium and technologies€¦ · cerbero goal 2 cross-layer...

CPS Summer School Alghero, 17 September 2018

CERBERO in a Nutshell Overview of Consortium and Technologies

Francesca Palumbo, Università degli Studi di Sassari (UNISS)

CERBERO Goal

2

Cross-layer modEl-based fRamework for multi-oBjective dEsign of Reconfigurable systems in unceRtain hybRid envirOnments

(CERBERO)

Integrated model-based methodology and initial framework for multi-objective design, incremental prototyping and continuous DevOps of Adaptive Cyber Physical Systems

• From (User Requirements) • SoS and System level • Application / Service level • Real Time Manager level • To Real Time Software and Hardware implementation

The Consortium: 12 partners from 7 countries

3

Started: January 1, 2017

Duration: 36 months

Components and Technology Providers

To build Adaptive Cyber Physical Systems

4

Started: January 1, 2017

Duration: 36 months

To build Adaptive Cyber Physical Systems

4

Started: January 1, 2017

Duration: 36 months

Ocean Space

Travelling

and evaluate by 3 use cases

5

Started: January 1, 2017

Duration: 36 months

CERBERO idea: from V to Ladder

6

CERBERO idea: from V to Ladder

6

CERBERO idea: from V to Ladder

6

CERBERO idea: from V to Ladder

7

CPS Issues - CERBERO Challenges

8

CPS Issues - CERBERO Challenges

8

CPS Issues

Modelling: beyond separation

of concerns

CPS Issues - CERBERO Challenges

8

CPS Issues

Modelling: beyond separation

of concerns

Design Reactive and Dynamic CPS:

Need for adaptivity support

CPS Issues - CERBERO Challenges

8

CPS Issues

Modelling: beyond separation

of concerns

Design Reactive and Dynamic CPS:

Need for adaptivity support

Design productivity gap:

model-based design automation

CPS Issues - CERBERO Challenges

8

CH1 … guaranteeing both functional and non-functional requirements by a holistic model-based cross-layer

engineering approach.

CH3 … contribute to standardization activities and open-

innovation initiatives … influence standards.

CPS Issues

Modelling: beyond separation

of concerns

Design Reactive and Dynamic CPS:

Need for adaptivity support

Design productivity gap:

model-based design automation

CH2 Reduce time-to-market, development efforts and cost of

ownership, of CPS and CPSoS.

CERBERO Challenges

Beyond SoA: CERBERO Activities to tackle CPS issues

• “BEYOND” (or better AID) SEPARATION OF CONCERNS: • Component-oriented approach: capable of cross-modelling,

optimization and analysis of hybrid systems. • Co-existence of time-continuum models and discrete ones. • Avoid transformations: combination of the different layers of

abstraction with unique models and tools or by means of appropriate semantics (CERBERO Cyber Physiscal Intermediate Format).

• BEYOND REQUIREMENTS ANALYSIS: • High level functional and non-functional requirements analysis

and continuous verification. • Generalization of requirements definition and management:

common Key Performance Indicators.

Modelling: beyond separation

of concerns

Design Reactive and Dynamic CPS:

Need for adaptivity support

Design productivity gap:

model-based design automation

Beyond SoA: CERBERO Activities to tackle CPS issues

• BEYOND SINGLE-LAYER: • Methodology for designing adaptive system architectures . • Environment, human and system triggers to enable

hardware/software (self-)reconfiguration. • Multi-layer adaptation: high-level self-adaptation manager,

which takes decision based on runtime KPI analysis.

Modelling: beyond separation

of concerns

Design Reactive and Dynamic CPS:

Need for adaptivity support

Design productivity gap:

model-based design automation

Beyond SoA: CERBERO Activities to tackle CPS issues

• BEYOND TOOL INTEGRATION: • Extensive support for runtime adaptive systems. • Incremental prototyping and verification, with system-in-the-loop

co-simulation capabilities. • Semantic integration of different design automation components.

Modelling: beyond separation

of concerns

Design Reactive and Dynamic CPS:

Need for adaptivity support

Design productivity gap:

model-based design automation

CERBERO R&D Core

12

WP3 CERBERO Model-Based

Development Approach for CPS WP5

CERBERO Continuous

Design & Operational Framework WP4

CERBERO Multi-Layer Runtime Adaptation

for CPS

CERBERO R&D Core

12

WP3 CERBERO Model-Based

Development Approach for CPS WP5

CERBERO Continuous

Design & Operational Framework WP4

CERBERO Multi-Layer Runtime Adaptation

for CPS

Modelling Productivity

Adaptivity

CERBERO R&D Core

12

WP3 CERBERO Model-Based

Development Approach for CPS WP5

CERBERO Continuous

Design & Operational Framework WP4

CERBERO Multi-Layer Runtime Adaptation

for CPS

CERBERO R&D Core

12

WP3 CERBERO Model-Based

Development Approach for CPS WP5

CERBERO Continuous

Design & Operational Framework WP4

CERBERO Multi-Layer Runtime Adaptation

for CPS

Adaptivity has to be captured at the model level. Models are meant to be

integrated in the self-adaptation engine, to master adaptivity.

CERBERO R&D Core

12

WP3 CERBERO Model-Based

Development Approach for CPS WP5

CERBERO Continuous

Design & Operational Framework WP4

CERBERO Multi-Layer Runtime Adaptation

for CPS

Adaptivity has to be captured at the model level. Models are meant to be

integrated in the self-adaptation engine, to master adaptivity.

Modelling methodologies &

models are the basis for the tools integration.

CERBERO R&D Core

12

WP3 CERBERO Model-Based

Development Approach for CPS WP5

CERBERO Continuous

Design & Operational Framework WP4

CERBERO Multi-Layer Runtime Adaptation

for CPS

Adaptivity has to be captured at the model level. Models are meant to be

integrated in the self-adaptation engine, to master adaptivity.

Runtime methodologies have to be supported by the design environment.

Tools aid designers to handle reconfiguration.

Modelling methodologies &

models are the basis for the tools integration.

• Raise the level of abstraction in CPS design, complementing separation of concerns.

CERBERO R&D Core: Modelling [WP3]

• Raise the level of abstraction in CPS design, complementing separation of concerns.

CERBERO R&D Core: Modelling [WP3]

13

Activities • Classified a relevant set of generic and re-

usable Key Performance Indicators (KPIs). • Surveyed Models of Computation to access

their features/limitations. • Definition of an integral approach for

modelling the different abstraction levels: CERBERO Intermediate Format (IF): • To make layers talk though KPIs. • To improve tools interoperability.

• Raise the level of abstraction in CPS design, complementing separation of concerns.

CERBERO R&D Core: Modelling [WP3]

Beyond SoA: • Advance Models of Computation • Extending semantic to improve expressiveness. • Supporting specification of non-functional KPIs.

• KPI-based methodology for: • Collecting the system requirement (including

long term ones) • Mapping the requirements to KPIs (in or out

library) • Continuously evaluating KPIs at run time and

adapt to changes

14

• Support heterogeneous system flexibility by defining multi-layer adaptation strategies implemented within a self-adaptation manager sensitive to human, environment and system triggers.

CERBERO R&D Core: Adaptivity [WP4]

15

• Support heterogeneous system flexibility by defining multi-layer adaptation strategies implemented within a self-adaptation manager sensitive to human, environment and system triggers.

CERBERO R&D Core: Adaptivity [WP4]

15

Activities • Common cross-layer view; • Single layer adaptation

strategies: HW adaptation strategies, proper monitors and engines, SW supervisors;

• Sensor/Information fusion techniques.

• Support heterogeneous system flexibility by defining multi-layer adaptation strategies implemented within a self-adaptation manager sensitive to human, environment and system triggers.

CERBERO R&D Core: Adaptivity [WP4]

16

Beyond SoA: • Address lack of toolchains

supporting adaptivity by developing suitable models/CAD for heterogeneous platforms in uncertain environments.

• Address lack of actual case studies.

CERBERO R&D Core: Design Productivity [WP5]

17

• Build an integrated design environment to model, explore, deploy and verify complex adaptive CPS.

Activities • Use-case to tool mapping.

• Design/enhancements of specific components.

• Integration • Semantic-based integration

strategy.

• Direct tool-to-tool integration.

CERBERO R&D Core: Design Productivity [WP5]

17

• Build an integrated design environment to model, explore, deploy and verify complex adaptive CPS.

Beyond SoA: • Address the lack of integrated

toolchains capable of: • Spanning across layers.

• Dealing with adaptivity and heterogeneity.

• Providing system in the loop co-simulation.

CERBERO R&D Core: Design Productivity [WP5]

18

• Build an integrated design environment to model, explore, deploy and verify complex adaptive CPS.

CERBERO Use Cases

19

Self-Healing System for Planetary Exploration: • Self-healing and self-adaptive embedded CPS processing systems capable of operating in such a

critical environment • Robotic arm and motors control for space vessel

CERBERO Use Cases

19

Self-Healing System for Planetary Exploration: • Self-healing and self-adaptive embedded CPS processing systems capable of operating in such a

critical environment • Robotic arm and motors control for space vessel

Ocean Monitoring: • Smart video-sensing unmanned vehicles with immersive environmental monitoring capabilities • Individual and fleet self-operation, power management and navigation • Data analysis and information fusion to enable smart adaptation strategies to address rapidly

changing environment conditions in order to obtain or maintain positions on sea and other missions objectives

CERBERO Use Cases

19

Self-Healing System for Planetary Exploration: • Self-healing and self-adaptive embedded CPS processing systems capable of operating in such a

critical environment • Robotic arm and motors control for space vessel

Ocean Monitoring: • Smart video-sensing unmanned vehicles with immersive environmental monitoring capabilities • Individual and fleet self-operation, power management and navigation • Data analysis and information fusion to enable smart adaptation strategies to address rapidly

changing environment conditions in order to obtain or maintain positions on sea and other missions objectives

Smart Travelling for Electric Vehicle: • Virtual Reality simulated environment for driver support in electric vehicles • Highly networked scenario composed of heterogeneous concurrent subsystems

• Electric Vehicle, Person possessing a only partially observable personal agenda, the Smart Energy Grid and the Smart Mobility that provides mobility-aware functionality (e.g. parking places, charge points, smart home, smart office, etc.)

• High degree of autonomy and support for adaptability, plus modelling and managing the distributed communication layers.

CERBERO @ Summer School

21

[18/09] Tutorial on Requirements Verification in CPS, Prof. Luca Pulina [Università degli Studi di Sassari], Prof. Armando Tacchella [Università degli Studi di Genova], Dr. Simone Vuotto [Università degli Studi di Sassari/Genova] [20/09] Tutorial on CPS Modeling and Exploration in The CERBERO Project, Dr. Michael Masin [IBM Research], Dr. Julio de Oliveira Filho [TNO] [20/09] Tutorial on Reconfigurable CPS Accelerators, Alfonso Rodriguez [Universidad Politecnica de Madrid] and Tiziana Fanni [Università degli Studi di Cagliari]

CERBERO @ Summer School

21

[18/09] Tutorial on Requirements Verification in CPS, Prof. Luca Pulina [Università degli Studi di Sassari], Prof. Armando Tacchella [Università degli Studi di Genova], Dr. Simone Vuotto [Università degli Studi di Sassari/Genova] [20/09] Tutorial on CPS Modeling and Exploration in The CERBERO Project, Dr. Michael Masin [IBM Research], Dr. Julio de Oliveira Filho [TNO] [20/09] Tutorial on Reconfigurable CPS Accelerators, Alfonso Rodriguez [Universidad Politecnica de Madrid] and Tiziana Fanni [Università degli Studi di Cagliari]

Meet our Demos Today @ Poster

Session!

Consortium Big Picture

22