START - Stochastic Real-Time Analysis of Embedded Software Systems
The aim of this project is to develop new models that, by using statistical inference techniques, are better suited for development of real industrial systems.
Description of the project
The real-time research community has over the years developed mature results in the area of real-time scheduling and timing analysis of embedded software systems. In parallel, engineers have created, developed, and evolved embedded software systems without complying with the academic real-time models required to do timing analysis. This gap between academic models and industrial practice is a key challenge in applying the (often very restrictive) academic models to real systems.
The aim of this project is to develop new models that, by using statistical inference techniques, are better suited for development of real industrial systems. Given these more expressive models, we will research and develop stochastic and simulation-based analysis of embedded software systems, providing software engineers with probabilistic measures of timing requirement violations for embedded software systems, in contrast to the more traditional and coarse grained pass/fail result provided by the conventional academic worst-case real-time analysis. The developed analysis will provide an important foundation for a broader adoption of sound timing analysis techniques for complex embedded software systems with real-time requirements.