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

Project Information

CERBERO

Grant agreement ID: 732105

DOI

10.3030/732105

Project closed

EC signature date 28 October 2016

Start date 1 January 2017

End date 29 February 2020

Funded under

• INDUSTRIAL LEADERSHIP - Leadership in enabling and industrial technologies - Information and Communication Technologies (ICT)

Total cost

€ 5 383 597,50

EU contribution

€ 4 996 460,00

Coordinated by

IBM ISRAEL - SCIENCE AND TECHNOLOGY LTD
Israel

Periodic Reporting for period 2 - CERBERO (Cross-layer modEl-based fRamework for multi-oBjective dEsign of Reconfigurable systems in unceRtain hybRid envirOnments)

Reporting period: 2018-07-01 to 2020-02-29

Summary of the context and overall objectives of the project

Cyber-Physical Systems (CPS) are complex systems composed of different interacting computing and physical entities that contribute concurrently to determine the behaviour of the system as a whole. Computing layer and physical environment are tightly bound; therefore, such systems need to adapt, prospectively and autonomously, to rapid changes in the environment and in the system itself. By their nature, CPS and System of Systems (CPSoS) usually operate in uncertain environments and should satisfy multiple concurring and, usually, competing requirements regarding affordability, performance, safety, security, sustainability, etc. As a result, the design of CPS and CPSoS becomes inherently difficult, challenging, and time consuming. Promising ways of addressing this can be to follow new development methodologies, such as NIST framework for CPS, and to use model-based design platforms both commercially available or academic. However, despite their big promise (considering the claimed enhancement of and the declared speed-up), the existing model-based frameworks are not as popular as it could be expected. Modelling, maintenance, and interoperability overhead, especially with heterogeneous models over several levels of abstraction, are not addressed in a satisfactory way. CERBERO intends to develop design methodology and address these challenges for a specific aspect of CPS – adaptivity. While deeply studied, there is no standard solution yet for adaptation and reconfiguration. In particular, self-reconfiguration and adaptation have been acknowledged as key features for CPS operators dealing with faults management, but existing design frameworks rarely address them. In order to focus CERBERO effort even more and evaluate the proposed framework and developed tools, CERBERO defined three use cases, targeting development of CPS in very different levels of abstraction and covering a wide spectrum of system features: (i) self-healing robotic arm for Planetary Exploration utilizing adaptability of heterogeneous embedded computing platform to improve computing robustness; (ii) video-sensing unmanned vehicles for Ocean Monitoring utilizing system level adaptivity for cost effective immersive environmental monitoring capabilities on-sea and subsea; and (iii) driver assistant for Smart Travelling of electric vehicles utilizing system of systems (SoS) level adaptivity including interaction with humans in an immersive simulation environment.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

The CERBERO consortiums vision is to extend and improve methods and tools to enable quick and cost effective design and deployment of interoperable, energy efficient, secure, and connected smart CPS and CPSoS devices. This is a long-term objective, which will continue long beyond the end of the project. To achieve the impact, CERBERO effort is divided to eight work packages (WP). WP1 provided smooth project progress with a lot of cooperation withing consortium.
In WP2 use case needs have been mapped to opportunities of using CERBERO technologies and technical requirements have been elicited based on use case needs and needs of other project stakeholders according to the generic methodology we developed.
WP3 formalized model semantics for adaptive CPS and defined requirements for intermediate system and metrics representation format and middleware for semantic models and KPI inter-operability between tools.
The main result of WP4 is the developed methodology for adaptation with all supporting definitions and its projection on hardware, software, sensing, information fusion, monitoring, and initial definition of hierarchical self-adaptation manager. The methodology addresses adaptation needs in all CERBERO use cases from mixed hardware adaptation to SoS level.
WP5 focuses on various aspects of productivity critical to usability and attractiveness of model-based design ranging from a prototype of optimization technologies increasing SoA for hybrid system in order of 10-100x trough initial prototype for building initial models and their check from natural language requirements and drastically improved accessibility of hardware adaptation to software/domain engineers to framework support of modular development using system-in-the-loop real time and look ahead simulations and prototypes of run time integration of tools with similar semantics and design time semantic integration of models and KPIs from different tools and, potentially, from different levels of abstraction using CERBERO Cyber-physical Intermediate Format (CIF) and initial CIF middleware Python API.
WP6 closes the loop with use case providers defining unified methodology for validation of CERBERO technologies in corresponding demonstrators.
Dissemination and Exploitation activities of WP7 and 8 included participation in leading conferences, organizing workshops and summer schools, participation in standardization commettee and defining exploitable items and business models.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

Time-to-market and design productivity can be largely improved by the adoption of comprehensive and automatic design toolchains. Several methodologies and frameworks are available for CPS engineering. These tools offer support for different aspects of modeling and design, and, usually, simulation. Only few support optimization and code generation. None of them intrinsically supports CPS adaptivity and reconfiguration or in-the-loop simulation. CERBERO design environment, whose components cover all the above-mentioned features, exploits cross-optimization of components model-based design to improve time-to-market and, in turn, productivity. Moreover, it is also capable of featuring support for heterogeneity that, despite a significant effort in this sense, is still an open issue. As an example, interoperability among AOW, DynAA and PREESM will allow the cross-optimization of heterogeneous multi-core nodes at the SoS level. PREESM will retrieve event-driven information from AOW or DynAA and combine it with dataflow information for cross-layer optimization (energy and latency aware) and for ensuring computation reliability (deadlock-freeness, livelockfreeness, memory management).
With respect to other framework-oriented EU projects, CERBERO can be compared with INTO-CPS (http://intocps.au.dk/) that aims at creating an integrated toolchain for comprehensive model-based design of CPS. It addresses modeling, design and verification, integrating existing industry strength tools, based centrally around Functional Mockup Interface compatible co-simulation. CERBERO, as well as INTO-CPS, leverages on a model-based approach and on a comprehensive design environment, but it offers also runtime support for adaptivity and cross-layer optimization. With respect to this latter, AOW, the core component in charge of cross-layer optimization in the CERBERO design environment, intends to extend Measure-based Continuous Linear Programming to support both linear and polynomial functions (more suitable for real-life problems), integrating objectives based on Robust Optimization with budgeted uncertainty and multi-objective Pareto-optimal hybrid solutions.

logowhite1.jpg