Application area: Healthcare Systems
Researcher in charge: Christos Kollintzas
Supervisor: Prof. Marc Geilen
Host: Eindhoven University of Technology, The Netherlands
Secondments: Philips Healthcare, The Netherlands
The main objective of this research is to develop techniques for managing cyber physical systems with frequent hardware platform changes. The project is conducted in collaboration between Eindhoven University of Technology and Philips Healthcare, with the Allura interventional X-Ray system being the cyber physical system under study.
The system’s complete functionality relies on software, from positioning the X-Ray beam to distributing video to monitors. This project focuses in the video distribution area of the system where real time requirements due to the nature of the system’s purpose as well as advances in technology demand for changes in the hardware platform with a relatively high frequency. These changes need to be tested for their compatibility with the existing software as well as for their impact in system performance. However, it is often observed that these hardware platforms may be deliverable of third-party suppliers resulting in unavailability to observe the component’s internal behaviour.
The researcher has started developing a tool for testing the performance of these third-party components as far as software is concerned. More specifically, the model of each component is derived by studying its behaviour at the interface, specifying expected interaction as well as timing behaviour. These interaction models in combination with timing traces from component execution are provided to Component Modelling and Analysis (ComMA) framework in order to create a verdict of whether the modelled behaviour is satisfied by the execution traces. ComMA is a framework developed in Philips Healthcare that supports model-based engineering of high-tech systems by formalising interface specifications. ESR’s current work is to create an integrated test system that implements specific use cases of the Allura’s video distribution system using these third-party components and decides upon their conformance to the specifications.
Future work will focus on extending the use cases implemented by the test system until full functionality of the video distribution system is covered. This way, problems that may occur during integration, as a result of software performance alterations between component versions, and may lead to a difficult to detect, unexpected, and erroneous system behaviour are avoided.
Furthermore, new concepts will be included in the model such as performance tolerance and model adaptation. More specifically, ESR will explore ways to derive estimated performance of new components and predict variations of it based on performance of current software versions. As far as model adaptation is concerned, the research will focus on reverse engineering of already existing interfaces, so that a model could be created from the traces of execution. This way, changes in the Philips implementation of use cases will generate an updated model to be provided to the test system instead of manually editing it.