From Post-Mortem to Prevention: Redefining “Invisible” Pedestrians Through ISO 26262 and Multi-Modal AI

Abstract

Nick Barua

Pedestrians in non-upright postures—those who have fallen due to medical emergencies, intoxication, or primary collisions— represent a significant yet underserved demographic in automotive safety research. Despite the proliferation of Advanced Driver- Assistance Systems (ADAS), forensic data suggests that “low-profile” humans remain largely invisible to standard monocular and LiDAR-based detection systems. This opinion piece argues for a paradigm shift in vehicle safety, moving from post-mortem forensic analysis to proactive prevention through the integration of ISO 26262 functional safety standards and multi-modal AI architectures. By leveraging the Advanced Falling Object Detection System (AFODS), which utilises YOLOv7-Tiny for spatial detection and MFCC-based audio classification for verification, detection accuracy for prone individuals can be significantly improved. Furthermore, the piece addresses the ethical implications of these systems, positing that the protection of the most vulnerable road users is a deontological necessity that transcends traditional “Trolley Problem” utilitarianism.

PDF