AI Driven Predictive Functional Safety Leveraging Key Performance Indicators in Autonomous Vehicles

As the automotive industry transitions towards Software-Defined Vehicles (SDVs) with centralized computing architectures, managing functional safety becomes increasingly complex. This paper examines how AI-driven decision systems can enhance predictive functional safety in autonomous vehicles especially with the architectural shift from traditional to zonal and centralized computing. A combined approach utilizing Predictive Artificial Intelligence (AI) and Key Performance Indicators (KPIs) is proposed to ensure functional safety and optimize system performance. KPIs track real-time performance of critical functions, serving as diagnostic tools during development and enabling AI-based predictive monitoring post-deployment. This approach facilitates proactive risk mitigation and continuous system optimization.