Safety-Oriented Optimization of Automotive E/E Architectures

The rapid development of advanced driver assistance systems (ADAS), connectivity, and autonomous driving is transforming vehicles into software-defined systems. As E/E architectures evolve toward centralized and zonal computing, the integration of hardware and software becomes increasingly complex, while adherence to the ISO 26262 functional safety standard remains essential for ensuring system reliability.

Dorsa Mohammad Zaheri presents a unified mixed-integer programming (MIP) framework for automated E/E architecture design. The methodology jointly optimizes task-to-hardware allocation, ASIL decomposition, message routing, and scheduling under ISO 26262 functional safety constraints. It supports multiple ISO 26262 design strategies and enables efficient tradeoff analysis among cost, performance, latency, and safety, delivering optimized solutions for mixed-criticality automotive systems.