Automatic Fault Tolerance Management based on Task Criticality for Real-Time Systems

초록

Safety-critical real-time systems, common in domains such as aerospace, defense, and autonomous driving, demand high reliability. However, conventional fault tolerance (FT) techniques are often applied uniformly, potentially compromising system stability by not prioritizing essential tasks. To address this, we propose a technique that automatically and dynamically assigns FT operations based on task criticality. It iteratively checks system schedulability while concentrating FT operations on critical tasks, ensuring equitable distribution. Then, an execution pattern is generated to distribute the FT overhead. Evaluations in a GR740 SIS environment supporting RTEMS 6 show that the proposed technique achieves higher schedulability than fixed-rate allocation approaches. © 2026 Copyright held by the owner/author(s).

키워드