Spatiotemporal-Aware Bit-Flip Injection on DNN-based Advanced Driver Assistance Systems (extended version)

TL;DR

This paper introduces a spatiotemporal-aware fault injection framework for DNNs in ADAS, identifying critical fault sites and timings to evaluate safety risks under realistic conditions.

Contribution

It proposes a novel framework combining spatial and temporal analysis to efficiently locate and trigger critical faults in DNN-based ADAS systems.

Findings

STAFI uncovers 29.56x more hazard-inducing faults than baseline methods.

The PMBS method effectively identifies critical network weight bits affecting driving behavior.

CFTI determines optimal fault timing to maximize safety impact.

Abstract

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.