CSR-EHS: Techniques for Assuring the Safety and Reliability of Physical Computing Systems and Applications to Medical Devices

该项目专注于开发可靠的物理计算系统（PCS），这是一类安全关键型嵌入式系统，通过传感器和执行器以重要方式与物理环境交互。目标应用域为医疗设备。随着医疗器械系统变得越来越复杂，现有的PCS设计技术已经不能有效地支持可靠和健壮的医疗器械系统的开发。新技术也将适用于其他领域，如航空电子系统、制造业和汽车控制器。该项目推进了形式化建模和模型驱动验证技术，使开发高度可靠的pc成为可能。该方法是通过显式建模故障将混合系统建模扩展到pc。然后将这些模型用于测试生成和运行时监视和检查。该项目还将通过开发静态和动态可靠性指标，为此类系统的严格认证过程奠定基础。最后，正在开发一个开放平台，用于创建和共享通用输液泵（GIP）的需求、设计、模型和测试套件的工件。该平台将培养一个对高度可靠的医疗设备感兴趣的社区，并最终使技术转化为实践。该项目的一个重要部分是与未来手术室的即插即用（PnP）开发人员的合作。在短期内，该项目将提高输液泵和其他计算机控制的医疗设备的质量，而长期影响将是在嵌入式系统工程中更广泛地使用正式方法。

The project concentrates on the development of reliable physicalcomputing systems (PCS), a class of safety-critical embedded systemsthat interact with the physical environment in significant ways viasensors and actuators. The target application domain is medicaldevices. The existing technology for PCS design does not effectivelysupport the development of reliable and robust medical device systemsas they become more complex. The new techniques will also haveapplicability to other domains such as avionic systems, manufacturing,and automotive controllers.The project advances formal modeling and model-driven validationtechniques to make it possible to develop highly dependable PCS. Theapproach is to extend hybrid systems modeling to PCS by explicitlymodeling failures. These models are then used for test generation andrun-time monitoring and checking. The project will also layfoundations for a rigorous certification process of such systems bydeveloping static and dynamic dependability metrics. Finally, an openplatform is being developed for creation and sharing of the artifactsof requirements, designs, models, test suites for generic infusionpumps (GIP). The platform will foster a community interested in highlydependable medical devices and ultimately enable transition of thetechnology into practice. A significant part of the project is thecollaboration with the developers of Plug-and-Play (PnP) for operatingroom of the future.In the short term, the project will improve the quality of infusionpumps and other computer-controlled medical devices, while long termimpact will be a wider use of formal approaches in theengineering of embedded systems.