Enhancing Functional Safety in Automotive AMS Circuits through Unsupervised Machine Learning

Overview

The framework injects faults across automotive analog and mixed-signal circuit levels and applies unsupervised feature clustering. A time-series extension examines early anomaly detection and latency as potential inputs to SoC safety mechanisms.