This presentation explores how automotive ECU timing performance can be measured, diagnosed, and optimized using a closed-loop engineering approach supported by local AI models. As vehicle E/E architectures evolve toward centralized, software-defined platforms, timing issues such as jitter, overruns, interrupt bursts, and multi-core contention become harder to identify through traditional methods. The session presents a practical framework for runtime measurement, anomaly detection, root-cause localization, optimization, and verification across AUTOSAR Classic and Adaptive systems, with particular relevance for integration quality and safety-critical vehicle software.

• How ECU timing issues emerge in software-defined vehicle architectures.
• How runtime measurement and local AI support anomaly detection and root-cause analysis.
• How closed-loop optimization improves integration quality and safety-critical timing evidence.