SynthSoft - Synthesizing Predictable Software for Distributed Embedded Systems
In this project we will bridge the sematic gap between (a) academic and industrial methods for timing modeling and analysis, and (b) industrial practices for model-based software development.
Project manager at MDU
We will do this by providing: (1) novel techniques for synthesis of predictable code from behavior- and component-models, (2) integration of timing-requirements modeling and timing-analysis tools in the design workflow. In this project, we target specifically the domain of distributed embedded real-time control systems, as represented by, e.g., automotive, aerospace and automation industries. A special focus will be on the automotive sector; a sector which is scientifically interesting due to the major improvement w.r.t. state-of-practice in software development over the last decade (including large-scale industrial adoption of techniques like component-based software engineering and model-based development) Concretely, this project will investigate how research oriented and/or standardized component models intended for the automotive domain (e.g. EAST-ADL, AUTOSAR) can be used together with component models actually used in industry today (Matlab/Simulink, Rubus Component Model) to provide both a functional description of the system as well as providing an analyzable and a resource efficient model of the run-time system, and how we can generate predictable and efficient code from these models.