Title: Bridging Gaps in ISO 26262 for Testing Machine Learning (ML) Systems

Author(s): Padma Iyenghar

Publication Event: Publication of Newsletter Volume 33 Nos 2 - May 2025

Publication Date: 2025-05-27

Resource URL: https://scsc.uk/r3123.pdf

Abstract:

Integrating Machine Learning (ML) into safety-critical automotive systems pre-sents new challenges. As automotive technologies evolve, standards like ISO 26262, which regulate the functional safety of electrical and electronic sys-tems in vehicles, must be updated to address the complexities of modern ML sys-tems. Drawing on a recent peer-reviewed IEEE Access publication co-authored with Emil Gracic and Gregor Pawelke, Padma Iyenghar outlines a systematic ap-proach to enhancing ISO 26262 by introducing ML-specific life cycle phases and testing methods to ensure the safety and reliability of ML-driven systems.