MESA - Metamodeling for Automation
July 01, 2005 to June 30, 2007
Software is now one of the key drivers of innovation in the automobile and will become even more important in the future. Yet automotive software must also meet high requirements, including those set by complex automotive systems architectures that use powerful networked computers. In this field traditional methods and tools for systems development can only be used to a limited extent and then not always efficiently. This means that automotive software and systems engineering will be one of the most important innovation-drivers in the coming years.
The MESA project (Metamodeling for Automation of Analysis and Development Methods for Software in Automobiles) is grounded in these issues of such importance for the automobile industry, and investigates how existing methods, architectures and processes for automotive software and systems engineering can be optimized through the use of innovative and proven techniques from software technology.
Below you can read more about the current challenges facing the automobile industry, project goals and deployed technologies. If you are interested in this or one of the related fields, please don’t hesitate to get in touch with us.
An increasing number of functions in moderns cars are now realized by electronic components, while the number of automotive functions controlled by software (like vehicle-interval radar, X-by-Wire, ESP etc.) is steadily rising.
Unfortunately modern automotive electronics are also one of the main causes of the problems that occur. Tackling the increasing system complexity of distributed software-based automotive functions will thus be a key development aspect for car manufacturers and suppliers alike with a decisive impact on their competitive edge. The problems associated with the development of embedded automotive systems are broadly similar to those found in other industry sectors:
- Development information is frequently only circulated in text form;
- The structure of development information and the relationships of single artifacts lack precise definition;
- Consistent implementation of any changes – and particularly those made in early on in the development cycle – is highly labor-intensive;
- Transparency and traceability of the individual development phases across the whole development cycle is very difficult to achieve;
- Only parts of the process are automated;
- The tools used are not integrated with one another.
The challenge of the MESA project is how to address these problems using modern methods of software technology, and especially those of model based development. At the same time heed must also be paid to the special framework conditions pertaining in the automobile industry which means that existing processes should be largely retained and existing development tools integrated in the new approach.
The project goals are:
- Metamodeling of development artifacts
Informed by new software development technologies, a concept is being developed whereby the semantics of development artifacts generated by the functional requirements analysis, systems architecture design and integration planning stages can be formalized as metamodels. The concept also analyses and factors in the special development processes particular to the automobile industry.
- Automotive-specific metamodels
Specific metamodels are produced for the above-mentioned development stages using the fully fledged concept and taking account of specific industry aspects and requirements.
- Metamodeling the interactions between artifacts
A metamodel-based concept is under development for the formalization of the types of relationships existing between such artifacts (consistency rules, dependencies, trace-relationships, analysis of change impact).
- Automotive-specific development guidelines
A concept is being developed and realized to enable the formalization and automatic inspection of development guidelines based on metamodels and the relationships between development artifacts they contain.
In addition to this, project findings will be used to develop a concept for integration in Volkswagen’s current development cycle, thus ensuring viability of commissioning under real life conditions.
The MESA projects makes consistent use of model based development techniques, in particular metamodeling and automated constraint checking. In a first stage relevant development information from a specific development stage is formalized in a MOF-compliant metamodel. Special generators are then used to produce “repositories for models” and integrate them with one another within a development infrastructure. This is based on the one hand on the repositories’ open standardized interfaces (CORBA, XML) and on the other on descriptions of the interactions of the development artifacts already contained in the metamodels. The special tools used by the automobile industry (such as Matlab/Simulink, DOORS) are connected with the respective repositories by adapters via open interfaces. Models can then be produced using traditional tools- yet at the end of the process all the relevant information is fed into the repository along with any additional information enabling transparency and traceability across various development stages.
OCL constraint checkers are also connected via the repositories’ open interfaces, enabling automatic inspection of guidelines previously formalized as OCL constraints.
The MESA project is funded through the Profit Program of the Investitionsbank Berlin and jointly implemented by FOKUS and Carmeq GmbH.