Improving Real-Time Performance Under Reliability Requirement Assurance in Automotive Electronic Systems

Abstract

The architecture of modern automotive electronic systems is composed of multiple distributed mixed-criticality functions. The effective utilization of such systems with heterogeneity is hindered by the scheduling of multifunctions. Minimizing the response time of functions (e.g., makespan) is an important optimization objective, and several scheduling algorithms for heterogeneous distributed architectures that optimize this objective have been proposed. However, the reliability requirements of functions of automotive systems are also an vital objective that needs to be optimized. Ignoring the reliability requirements of the functions during the scheduling process can cause catastrophic consequences, especially in automotive electronic systems. We propose a heuristic scheduling algorithm called real-time aware scheduling under reliability requirement assurance (RTRRA). The proposed algorithm meets the reliability requirement of each function by reliability transfer method technique and reduces the high-criticality functions' deadline miss ratio (DMR) by applying resource reallocation policy. Our results indicate that the RTRRA can reduce the response time of critical functions and meet the reliability requirements of all functions compared with the state-of-the-art algorithms.

Keywords

• Automotive functional safety
• bicriteria optimization
• mixed-criticality
• reliability requirement