SPEculative and Exploratory Design in Systems Engineering Project

The four-year SPEEDS (SPEculative and Exploratory Design in Systems Engineering) project, funded under the European Union’s 6th Framework Programme, has come to a highly successful conclusion. The SPEEDS project has resulted in the definition of a novel end-to-end design methodology, process and tool environment for model-based safety-critical embedded systems that significantly improves design quality while reducing both design cycle times and costs.

Pre-SPEEDS ideal development paradigm diagram

“A unique collaboration between embedded systems researchers, developers and design tool vendors, the successful conclusion of the SPEEDS project will substantially improve the competitiveness of the European embedded systems industry,” said Gert Döhmen, Coordinator of the SPEEDS Consortium and Head of Processes and Methods, Avionics Platforms and Electrical Systems at Airbus Deutschland. “SPEEDS provides the basis for the European embedded systems industry to evolve from model-based design of hardware and software systems towards the design and construction of integrated, component-based complete virtual system models.”

SPEEDS achieves this by delivering a multi-layered contract-based design and analysis methodology that formalizes the concept of Assumptions and Promises inherent within contracts. A new ‘controlled speculative design process’ – formalized in the SPEEDS Process Advisor toolset – facilitates trade-off studies through the evaluation of different architectures, enables right-first-time design and minimizes the risk in adopting concurrent design practices.

Building on SysML, the concept of Heterogeneous Rich Components (HRC), in the form of the tool-independent SPEEDS meta-model, underpins the SPEEDS methodology as the basis for model-driven engineering design. The SPEEDS meta-model is expressive enough to cover the complete development cycle from high-level specifications to design models as well as address both functional and non-functional design aspects. The HRC components are characterized by formal contracts allowing the use of various analysis techniques, including the new concept of hosted simulation, for design validation in the earliest design stages. The SPEEDS meta-model has also been constructed to seamlessly extend the capabilities of existing industry-specific SysML-based meta-models such as AUTOSAR and AADL.

Supporting the SPEEDS methodology is the SPEEDS Tool Architecture. The tool-independence of the SPEEDS meta-model enables the easy integration of any commercially available embedded systems design tool as it avoids the need for file exchange. Design tool integration is built on the defacto Eclipse standard using the SPEEDS Bus for which the API has been published.

“The SPEEDS design framework is now a reality,” said Roger Dierks, Manager Exploitation & Dissemination, SPEEDS and Sales Manager of Atego Deutschland. “Its adoption will further, significantly drive improvements in time-to-market and help achieve the speedier deployment of embedded systems designs.”

The SPEEDS methodology is also real-world proven having been successfully implemented in pilot projects by the SPEEDS user partners utilizing SPEEDS-compliant tools from the project’s vendor partners. Over 20 large enterprises are now planning to exploit the methodology and concepts as part of their systems design processes.

“The exploitation of the SPEEDS project results is rapidly gaining momentum through implementation of integrated tool environments,” said Dr. Michael Winokur, Technical Manager of the SPEEDS project and Corporate Director of Engineering and Development at Israel Aerospace Industries (IAI). “This is being achieved through the insertion of the multi-layered contract-based design methodology and its underlying HRC meta-model in the industrial partners’ engineering design processes, coupled with the tight support of the tool vendors involved in the project who are keen to commercialize the results by incorporating them in their tool suites.”

For developers wishing to explore the potential benefits of the SPEEDS methodology, a variety of materials have been published and are available for download. These include the SPEEDS meta-model and implementation guide, HRC Compiler, Contract Specification Language, SPEEDS Lessons and Best Practices guide and a SPEEDS training kit. The SPEEDS consortium is also working towards the publication of a book that will make the SPEEDS methodology and design process available for widespread industrial use, university courses and technical training for both end users and technology providers.

About the SPEEDS consortium
The SPEEDS consortium comprises the following companies and research institutions: Airbus, Atego (Extessy), Carmeq, EADS, Esterel Technologies, Dassault Systèmes (Geensoft), IBM Israel, Institut National de Récherche en Informatique et en Automatique (INRIA), Israel Aerospace Industries (IAI), Knorr Bremse Fekrendszerek, Kuratorium OFFIS, Magna Powertrain Engineering Center, Parades Geie, Robert Bosch, Saab and Université Joseph Fourier Grenoble. The SPEEDS initiative is a concerted effort to define a standard end-to-end framework for the implementation of innovative, next generation concepts, methodologies, processes, technologies and tools for the design of embedded systems. It is intended to significantly improve Europe’s performance and competitiveness in embedded systems design in key, safety-critical industry sectors such as automotive, avionics and industrial control, in what are increasingly becoming fiercely competitive markets. Its aim is also to provide an environment that will not only foster greater collaboration between European companies of all sizes involved in the embedded systems arena but also significantly widen access for smaller companies to leading edge tools and techniques. The initiative as a whole will enable the European embedded systems industry to seamlessly evolve from model-based design of hardware/software systems towards an integrated component-based construction of complete, virtual system models.