Integrating reliability and timing analysis of CAN-based systems

Abstract

The paper outlines and illustrates a reliability analysis method developed with a focus on CAN based automotive systems. The method considers the effect of faults on schedulability analysis and its impact on the reliability estimation of the system, and attempts to integrate both to aid system developers. We illustrate the method by modeling a simple distributed antilock braking system, showing that even in cases where the worst-case analysis deem the system unschedulable, it may be proven to satisfy its timing requirements with a sufficiently high probability. From a reliability and cost perspective, the paper underlines the tradeoffs between timing guarantees, the level of hardware and software faults, and per-unit cost.