In safety-critical embedded systems, functionalities with different criticalities are often supported upon shared hardware platforms, to enable size, weight, and power (SWaP) reductions. For example, in an automotive system, highly critical vehicle-control functions and less critical driver-alert functions might be hosted on the same hardware platform. While such platform integration may be cost-effective, care must be taken to isolate failures of less-critical components so that more-critical components are not affected. This project is directed at developing techniques for ensuring such isolation, particularly on hardware platforms that utilize multicore processors.
Specific goals in this project include devising new models for representing mixed-criticality systems, new metrics for quantifying the effectiveness of techniques for designing such systems (particularly from a SWaP perspective), and new methods for performing resource allocation and scheduling in integrated architectures that support mixed-criticality systems. The overall objective is to use these new models, metrics, and methods, to obtain new tools, techniques, and methodologies for deriving mixed-criticality system designs that are both correct by construction and implementable in a resource-efficient manner.
The investigators in this project collaborate extensively with partners in the avionics and automotive industries. Continued collaborations are expected to enable the results of this project to push research on mixed-criticality systems forward to address national industrial needs. This project is also expected to produce new open-source software and tools, new course content, and public outreach through participation in UNC's demo program (which is aimed at K-12 students, college students, and school teachers) and lectures and seminars by the investigators at national and international forums.