다중 주기의 혼합 임계 실시간 태스크 모델 연구

Abstract

This study introduces a novel task model for mixed-criticality real-time tasks with various periods, leveraging the Earliest Deadline First (EDF) scheduling algorithm. In the proposed model, a task might have different periods for the system mixed-criticality. While a high-criticality task requires more frequent execution in the high system mode, while a low-criticality task runs less frequently in the high mode. Since the traditional mixed-criticality real-time systems assume the ignorance of low-criticality tasks in the high mode, the proposed model is more suitable for designing the intricate scheduling requirements of complex systems, particularly in autonomous driving scenarios. Through comprehensive simulations, encompassing utilization intervals, task periodicity, and criticality level ratios, the study evaluates the scheduling efficacy of the proposed model. Results demonstrate the model's robustness in congested situations, surpassing conventional utilization limits. In comparison to existing mixed-criticality models, our proposed approach exhibits superior performance, showcasing its potential applicability to intricate systems such as autonomous vehicles, contributing to enhanced adaptability and efficiency in real-time task scheduling.