Spatiotemporal-Aware Bit-Flip Injection on DNN-based Advanced Driver Assistance Systems
For ADAS safety engineers, this provides a method to find DNN vulnerabilities that could cause accidents, but it is an incremental improvement over existing fault injection techniques.
The paper introduces STAFI, a framework that efficiently locates critical fault sites in DNNs for ADAS by identifying which weight bits and timing of faults cause the largest safety hazards. Experiments show STAFI uncovers 29.56x more hazard-inducing critical faults than the strongest baseline.
Modern advanced driver assistance systems (ADAS) rely on deep neural networks (DNNs) for perception and planning. Since DNNs' parameters reside in DRAM during inference, bit flips caused by cosmic radiation or low-voltage operation may corrupt DNN computations, distort driving decisions, and lead to real-world incidents. This paper presents a SpatioTemporal-Aware Fault Injection (STAFI) framework to locate critical fault sites in DNNs for ADAS efficiently. Spatially, we propose a Progressive Metric-guided Bit Search (PMBS) that efficiently identifies critical network weight bits whose corruption causes the largest deviations in driving behavior (e.g., unintended acceleration or steering). Furthermore, we develop a Critical Fault Time Identification (CFTI) mechanism that determines when to trigger these faults, taking into account the context of real-time systems and environmental states, to maximize the safety impact. Experiments on DNNs for a production ADAS demonstrate that STAFI uncovers 29.56x more hazard-inducing critical faults than the strongest baseline.