Published on Mon Jul 01 2019

ML-based Fault Injection for Autonomous Vehicles: A Case for Bayesian Fault Injection

Saurabh Jha, Subho S. Banerjee, Timothy Tsai, Siva K. S. Hari, Michael B. Sullivan, Zbigniew T. Kalbarczyk, Stephen W. Keckler, Ravishankar K. Iyer

DriveFI is a machine-learning-based fault injection engine. It can mine situations and faults that maximally impact AV safety. DriveFI found 561 safety-critical faults in less than 4 hours. In comparison, random injection experiments over several weeks could not find any.

0
0
0
Abstract

The safety and resilience of fully autonomous vehicles (AVs) are of significant concern, as exemplified by several headline-making accidents. While AV development today involves verification, validation, and testing, end-to-end assessment of AV systems under accidental faults in realistic driving scenarios has been largely unexplored. This paper presents DriveFI, a machine learning-based fault injection engine, which can mine situations and faults that maximally impact AV safety, as demonstrated on two industry-grade AV technology stacks (from NVIDIA and Baidu). For example, DriveFI found 561 safety-critical faults in less than 4 hours. In comparison, random injection experiments executed over several weeks could not find any safety-critical faults

Wed Jun 09 2021
Artificial Intelligence
Practical Machine Learning Safety: A Survey and Primer
0
0
0
Wed Oct 31 2018
Machine Learning
Scalable End-to-End Autonomous Vehicle Testing via Rare-event Simulation
Real-world testing, the evaluation environment, places the public in danger, and, due to the rare nature of accidents, will require Billions of miles in order to statistically validate performance claims. We implement a simulation framework that can test an entire modern autonomous driving system.
0
0
0
Sun Dec 08 2019
Machine Learning
Efficient Black-box Assessment of Autonomous Vehicle Safety
Real-world testing, the evaluation method, is dangerous to the public. We implement a simulation testing framework that evaluates an entire modern AV system as a black box. Using this framework, we conduct the first independent evaluation of a full-stack commercial AV system.
0
0
0
Wed Aug 19 2020
Artificial Intelligence
Assessing Safety-Critical Systems from Operational Testing: A Study on Autonomous Vehicles
Autonomous Vehicles (AVs) are making their debut on public roads. Methods for assessing whether an AV is safe enough are urgently needed. We apply new theorems extending Conservative Bayesian Inference (CBI) to AVs.
0
0
0
Fri Apr 24 2020
Machine Learning
ML-driven Malware that Targets AV Safety
Security attacks that violate safety constraints and cause accidents are a significant deterrent to achieving public mistrust in AVs. Creating a security hazard that results in a severe safety compromise is compelling from an attacker's perspective. For example, the smart purposefullymalware caused 33X more forced emergency braking than random attacks
0
0
0
Wed Apr 18 2018
Artificial Intelligence
Simulation-based Adversarial Test Generation for Autonomous Vehicles with Machine Learning Components
Autonomous driving systems are expected to be deployed at a large scale in the near future. Despite this, there is a lack of agreement on appropriate methods to test, debug, and certify these systems. We present a testing framework that is compatible with test case generation andautomatic falsification methods.
0
0
0