Utilization-Based Safety Aware Scheduling of Mixed-Criticality Real-Time System With Gradual Performance Degradation

Mixed-criticality (MC) systems have become indispensable in managing diverse tasks within unified computational environments, particularly in complex embedded systems like those in the automotive and avionics industries. Traditional models often discard low-criticality tasks abruptly during mode switches, while alternative strategies propose gradual adjustments to maintain balance. However, these existing approaches can compromise system performance and safety. This study presents SEMCAP, a novel scheduling method for MC systems that dynamically adjusts task periods to optimize resource allocation while ensuring safety and meeting timing requirements. The study introduces schedulability conditions enabling real-time evaluation of task feasibility. Extensive experiments demonstrate that SEMCAP improves the performance of low-criticality tasks compared to existing strategies, highlighting its superiority in managing mixed-criticality scenarios and achieving optimal resource utilization.