The cluster activities are relevant for industrial sectors in which a major challenge is the need for mastering system integration of complex heterogeneous embedded systems. Several activities focus particularly on the automotive and aeronautics sectors.
- Aeronautics
This sector is faced with the challenge of Integrated Modular Avionics (IMA), which drastically changes the OEM/supplier relations. Integration will occur at the level of functions, not any more at the level of packaged hardware modules and devices. Therefore, constructors are faced with the need of mastering system integration at all levels of the design process (from requirements to hardware). This move will drastically impact how certification will be performed in the future. - Automobile
The move is similar to that in aeronautics, the changes being in fact much more rapid and drastic – within a few years, the OEM/supplier chain will be entirely reconfigurated. Added value, for constructors, will move to completely different components of the car, namely those mostly contributing to building the “concept” and “personality” of each different car. Sharing platforms with competitors is now the trend, as shown by the Autosar initiative.
Both the above sectors are currently driving the innovation in the area of system level integration of complex heterogeneous embedded systems, for the following reasons:
Complex hardware systems or System on Chip (SoC) design does not involve the same degree of heterogeneity since plant and physical process for control are not part of the problem.
Research related to real-time components in the area of automation is still mostly academic, with the industrial move being hampered by the lack of agreed formal bases for the IEC 61131-3 and IEC 61499 standards for distributed control systems.
Regarding the area of large information systems (such as military systems, air traffic control systems, telecommunication network systems…), the main focus is on component based software development in general, with a lesser emphasis on real-time aspects.
Both the model- and the component-based development approaches allow integration problems to be handled in the earlier phases of system design. Component properties that have global system impact, notably properties of timing and resource consumption, can be specified in interfaces in such a way that global resource usage can be predicted a priori, avoiding hard problems in system integration. The research goals of the cluster address these challenges.
Wider adoption of model-based approaches will be supported by standards for modelling of Real-Time Embedded Systems, as well as by the availability of tool chains for design, transformation and analysis that are based on such standards. These issues are addressed in the activities on Components Platform for Component Modelling and Verification and Development of UML for Real-Time Embedded Systems.
Composition of models developed for different parts and viewpoints needs techniques for dealing with heterogeneity in design flows. This includes mixing different styles of scheduling policies (e.g., Event-Triggered and Time-Triggered), as well as mixing different Models of Computation. These issues are addressed in the activities on Forums with specific industrial sectors and Seeding new research directions. The results of these activities will in turn help to improve future model transformation and analysis tools.
Further technical discussions are needed in order to understand how these problems can be handled in the context of industrial system development.