Theoretical Modeling Research on Mixed Criticality System for Central Computing Platforms

To address the conflict between the resource waste caused by the necessity for safety isolation and the efficient resource utilization required by the principle of sharing when deploying tasks of different criticality on a central computing platform, an Automotive Multi Mode Mixed Criticality System (AMMMCS) and an Automotive Adaptive Mixed Criticality System (AAMCS), along with their respective degradation strategies, were proposed based on mixed-criticality theory. The AMMMCS was arranged for low-criticality tasks to adopt a relatively proactive scheduling during the high-criticality task executions for driving safety. The AAMCS was designed to ensure the execution of all tasks with maximize efficiency as long as vehicle control tasks within the threshold. The SMC-NO, AMC, AMMMCS, and AAMCS models were arranged on the FreeRTOS system and the Hardware-in-the-Loop experiments were carried out. The results show that the proposed methods can significantly improve task fulfillment rates and processor resource utilization.