Deepbillboard: Systematic physical-world testing of autonomous driving systems

Zhou，Husheng1; Li，Wei2; Kong，Zelun1; Guo，Junfeng1; Zhang，Yuqun1; Yu，Bei3; Zhang，Lingming1; Liu，Cong1

10.1145/3377811.3380422

2020-06-27

0270-5257

978-1-7281-6519-6

Proceedings - International Conference on Software Engineering

347-358

5-11 Oct. 2020

Seoul, Korea (South)

Deep Neural Networks (DNNs) have been widely applied in autonomous systems such as self-driving vehicles. Recently, DNN testing has been intensively studied to automatically generate adversarial examples, which inject small-magnitude perturbations into inputs to test DNNs under extreme situations. While existing testing techniques prove to be effective, particularly for autonomous driving, they mostly focus on generating digital adversarial perturbations, e.g., changing image pixels, which may never happen in the physical world. Thus, there is a critical missing piece in the literature on autonomous driving testing: understanding and exploiting both digital and physical adversarial perturbation generation for impacting steering decisions. In this paper, we propose a systematic physicalworld testing approach, namely DeepBillboard, targeting at a quite common and practical driving scenario: drive-by billboards. Deep- Billboard is capable of generating a robust and resilient printable adversarial billboard test, which works under dynamic changing driving conditions including viewing angle, distance, and lighting. The objective is to maximize the possibility, degree, and duration of the steering-angle errors of an autonomous vehicle driving by our generated adversarial billboard. We have extensively evaluated the efficacy and robustness of DeepBillboard by conducting both experiments with digital perturbations and physical-world case studies. The digital experimental results show that DeepBillboard is effective for various steering models and scenes. Furthermore, the physical case studies demonstrate that DeepBillboard is sufficiently robust and resilient for generating physical-world adversarial billboard tests for real-world driving under various weather conditions, being able to mislead the average steering angle error up to 26.44 degrees. To the best of our knowledge, this is the first study demonstrating the possibility of generating realistic and continuous physical-world tests for practical autonomous driving systems; moreover, Deep- Billboard can be directly generalized to a variety of other physical entities/surfaces along the curbside, e.g., a graffiti painted on a wall.

Autonomous Driving Adversarial Machine Learning Steering Model Testing

其他

英语

EI

20204409433186

Testing ; Automobile steering equipment ; Deep neural networks ; Steering

Highway Transportation:432 ; Ergonomics and Human Factors Engineering:461.4 ; Automobile and Smaller Vehicle Components:662.4 ; Robot Applications:731.6

2-s2.0-85092704737

Scopus

1.University of Texas at Dallas,Dallas,United States
2.Southern University of Science and Technology,Shenzhen,China
3.Chinese University of Hong Kong,Hong Kong,Hong Kong

Zhou，Husheng,Li，Wei,Kong，Zelun,et al. Deepbillboard: Systematic physical-world testing of autonomous driving systems[C],2020:347-358.