Model-driven system development for distributed fuel management in avionics

Model-based engineering for control software architectures reduces risks and costs in development processes of embedded systems. Developers who follow this methodological approach are able to evaluate the systemwide analysis impact of architectural choices and increase confidence through early verification/validation of the model assumptions. This design methodology is well suitable for a European project (SmartFuel) that proposes a distributed control strategy for aircraft fuel systems. During the first SmartFuel phase, the verification and validation of the aforementioned avionics system was successfully carried out through prototype-based simulations. The future second project phase develops models and tools to analyze, synthesize, and preverify/prevalidate the avionics system architecture at early development stages. This paper addresses the latter project phase by focusing on the prediction of system quality characteristics and the automatic code generation for the system components. It presents a software development framework to deal with the avionics system specifications by means of architectural representations that facilitate the early evaluation of the system. The proposed modeling methodology is evaluated in simulations and prototypes, and it is compared with the results obtained from the first project phase. The experimental outcomes from two case studies (helicopter and airplane fuel systems) are also presented.