Reconstruction-Based Adversarial Attack Detection in Vision-Based Autonomous Driving Systems

被引:9
|
作者
Hussain, Manzoor [1 ]
Hong, Jang-Eui [1 ]
机构
[1] Chungbuk Natl Univ, Dept Comp Sci, Software Intelligence Engn Lab, Cheongju 28644, South Korea
来源
关键词
deep learning; adversarial attacks; robustness; safety; autonomous vehicles; autoencoders; PERTURBATIONS; RESISTANT; VEHICLES; SAFETY;
D O I
10.3390/make5040080
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The perception system is a safety-critical component that directly impacts the overall safety of autonomous driving systems (ADSs). It is imperative to ensure the robustness of the deep-learning model used in the perception system. However, studies have shown that these models are highly vulnerable to the adversarial perturbation of input data. The existing works mainly focused on studying the impact of these adversarial attacks on classification rather than regression models. Therefore, this paper first introduces two generalized methods for perturbation-based attacks: (1) We used naturally occurring noises to create perturbations in the input data. (2) We introduce a modified square, HopSkipJump, and decision-based/boundary attack to attack the regression models used in ADSs. Then, we propose a deep-autoencoder-based adversarial attack detector. In addition to offline evaluation metrics (e.g., F1 score and precision, etc.), we introduce an online evaluation framework to evaluate the robustness of the model under attack. The framework considers the reconstruction loss of the deep autoencoder that validates the robustness of the models under attack in an end-to-end fashion at runtime. Our experimental results showed that the proposed adversarial attack detector could detect square, HopSkipJump, and decision-based/boundary attacks with a true positive rate (TPR) of 93%.
引用
收藏
页码:1589 / 1611
页数:23
相关论文
共 50 条
  • [41] A comparative study of vision-based lateral control strategies for autonomous highway driving
    Kosecka, J
    Blasi, R
    Taylor, CJ
    Malik, J
    1998 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION, VOLS 1-4, 1998, : 1903 - 1908
  • [42] Challenges of Designing Computer Vision-based Pedestrian Detector for Supporting Autonomous Driving
    Sun, Peng
    Boukerche, Azzedine
    2019 IEEE 16TH INTERNATIONAL CONFERENCE ON MOBILE AD HOC AND SMART SYSTEMS (MASS 2019), 2019, : 28 - 36
  • [43] GRI: General Reinforced Imitation and Its Application to Vision-Based Autonomous Driving
    Chekroun, Raphael
    Toromanoff, Marin
    Hornauer, Sascha
    Moutarde, Fabien
    ROBOTICS, 2023, 12 (05)
  • [44] A color vision-based lane tracking system for autonomous driving on unmarked roads
    Sotelo, MA
    Rodriguez, FJ
    Magdalena, L
    Bergasa, LM
    Boquete, L
    AUTONOMOUS ROBOTS, 2004, 16 (01) : 95 - 116
  • [45] AN AUTONOMOUS VISION-BASED MOBILE ROBOT
    BAUMGARTNER, ET
    SKAAR, SB
    IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 1994, 39 (03) : 493 - 502
  • [46] Vision-based autonomous soccer robots
    Khessal, NO
    Naing, MY
    Hwee, ENB
    Oo, PS
    Antony, LHS
    IEEE 2000 TENCON PROCEEDINGS, VOLS I-III: INTELLIGENT SYSTEMS AND TECHNOLOGIES FOR THE NEW MILLENNIUM, 2000, : 207 - 212
  • [47] A Vision-Based Approach for Autonomous Landing
    Cabrera-Poncel, Aldrich A.
    Martinez-Carranza, Jose
    2017 WORKSHOP ON RESEARCH, EDUCATION AND DEVELOPMENT OF UNMANNED AERIAL SYSTEMS (RED-UAS), 2017, : 126 - 131
  • [48] Vision-based robotic convoy driving
    Schneiderman, H
    Nashman, R
    Wavering, A
    Lumia, R
    MACHINE VISION AND APPLICATIONS, 1995, 8 (06) : 359 - 364
  • [49] Obstacle Classification and Detection for Vision Based Navigation for Autonomous Driving
    Deepika, N.
    Variyar, Sajith V. V.
    2017 INTERNATIONAL CONFERENCE ON ADVANCES IN COMPUTING, COMMUNICATIONS AND INFORMATICS (ICACCI), 2017, : 2092 - 2097
  • [50] Demo: Security of Camera-based Perception for Autonomous Driving under Adversarial Attack
    DiPalma, Christopher
    Wang, Ningfei
    Sato, Takami
    Chen, Qi Alfred
    2021 IEEE SYMPOSIUM ON SECURITY AND PRIVACY WORKSHOPS (SPW 2021), 2021, : 243 - 243