Real-Time Scheduling for Mixed-Criticality Systems in the Automotive Industry
Junghwan Lee, Myung-Jun Kim · Journal of Computing Science and Engineering · 2020
Currently, mixed-criticality systems (MCSs) are rapidly adopted by the automotive industry, with the shift in electricalelectronic architecture from federated to integrated design to reduce developmental costs, pull in the development schedule, and easy reconfiguration of the system with service-oriented architecture. Several studies have been based on Vestal’s original MCS model, in which the criticality modes are the same as criticality levels. However, the MCS model does not fit the automotive industry or the safety perspective. In this study, we identify the divergence of theory and automotive practice for real-time MCS. We also propose a generalized MCS model close to industry practice and a priority assignment algorithm along with schedulability analysis for both online and offline phases. Further, we present a practical example of memory partition and decomposition tasks based on AUTomotive Open System Architecture (AUTOSAR). The proposed design is currently being developed for battery management systems of electric and plug-in hybrid electric vehicles.