AI-Enabled Functional Safety: Sensor-Centric Strategies from ADAS to Autonomous Driving

Session details:

Functional Safety (FuSa) has become a defining pillar in the evolution of Advanced Driver Assistance Systems (ADAS) toward fully autonomous driving. As vehicles rely on increasingly complex sensor ecosystems—cameras, radar, lidar, lighting, and battery management systems—the challenge lies in ensuring that these components not only perform reliably but also comply with ISO 26262 and evolving global safety standards.

This session explores AI-enabled functional safety strategies that strengthen sensor reliability and fault tolerance across critical automotive domains. Real-world case studies highlight how sensor-centric FuSa design reduces collisions by up to 30%, improves pedestrian detection in low-light conditions, and prevents up to 95% of EV thermal runaway incidents through safety-certified battery management systems. We will examine how AI augments traditional hazard analysis and risk assessment (HARA), enabling predictive fault detection, adaptive safety monitoring, and enhanced explainability of safety-critical decisions.

The presentation will also address research gaps and future directions, including the integration of Safety of the Intended Functionality (SOTIF), AI transparency, and the convergence of FuSa with cybersecurity frameworks. Attendees will gain insights into sensor-focused design strategies that balance innovation with compliance, accelerating the safe deployment of autonomous mobility.