The various initiatives that have investigated the software issues in projects (mainly the fact that software is always on the critical path) have demonstrated that a major concern was the weakness of the link between the avionics system and the software. The systems are more and more complex. Their definition comes later and later in the project. The software requirements are never stable. The software schedule becomes shorter and shorter. These changes mean that software engineers have to implement requirements which are still moving when the software is already in AIT.
One solution is to improve the avionics system-software co-engineering activities, by giving system and software teams (i) interleaved and consistent processes and (ii) common requirements tools. The best way to have a common understanding of requirements is to build a model. ESA has investigated Avionics Architectural Description Languages and is currently concentrating its efforts on AADL.
AADL is being standardized by the Society of Automotive Engineers (SAE) in its AS-2C Architecture Description Language Subcommittee. AADL has used the Honeywell Meta-H specification as major input. ESA is a voting member of the committee, along with major space and aircraft industry participants.
Automated System and Software Engineering for Real-Time applications (ASSERT)
ASSERT has a collaborative web site for the members and a public web site. ASSERT combines the use of AADL to describe the avionics and its properties with a proof based system engineering approach, takes into account the dependability properties, automates the subsequent phases of the life cycle, delivers supporting tools and processes, and disseminates the results.
The avionics system-software co-engineering activities are performed as a collaborative effort by the ESTEC entities in responsible for software, data handling and control.
Hardware Software Co-design
Another dimension of the System-Software co-engineering is the link with the functionality implemented in microelectronics. From the system models, a transformation is possible to microelectronic languages (examples: systemC or VHDL) for those functions that will be implemented in microelectronics (ASIC, FPGA, SoC). This includes the trade-off of the partitioning of the system functions into software or microelectronics implementations.
Correctness, Modeling and Performance of Aerospace systems
The COMPASS Project is an international research project for developing a theoretical and technological basis for the system-software co-engineering approach focusing on a coherent set of specification and analysis techniques for evaluation of system-level correctness, safety, dependability and performability of on-board computer-based aerospace systems.
In this project a model-based approach to system-software co-engineering is taken, specifically tailored to critical on-board systems for the space domain. For this, the consortium develops an integrated toolset based on a newly developed modelling language.
The envisioned COMPASS integrated platform toolset comprises of several model validation and reduction tools, all integrated together, including in particular a Formal Safety Analysis Platform, a Model Based Planner, a Model Reward Model Checker, a Requirements Analysis Tool
, and NuSMV
Last update: 3 October 2011