A Caltech Library Service

Robust Trajectory Prediction against Adversarial Attacks

Cao, Yulong and Xu, Danfei and Weng, Xinshuo and Mao, Zhuoqing and Anandkumar, Anima and Xiao, Chaowei and Pavone, Marco (2022) Robust Trajectory Prediction against Adversarial Attacks. . (Unpublished)

Full text is not posted in this repository. Consult Related URLs below.

Use this Persistent URL to link to this item:


Trajectory prediction using deep neural networks (DNNs) is an essential component of autonomous driving (AD) systems. However, these methods are vulnerable to adversarial attacks, leading to serious consequences such as collisions. In this work, we identify two key ingredients to defend trajectory prediction models against adversarial attacks including (1) designing effective adversarial training methods and (2) adding domain-specific data augmentation to mitigate the performance degradation on clean data. We demonstrate that our method is able to improve the performance by 46% on adversarial data and at the cost of only 3% performance degradation on clean data, compared to the model trained with clean data. Additionally, compared to existing robust methods, our method can improve performance by 21% on adversarial examples and 9% on clean data. Our robust model is evaluated with a planner to study its downstream impacts. We demonstrate that our model can significantly reduce the severe accident rates (e.g., collisions and off-road driving).

Item Type:Report or Paper (Discussion Paper)
Related URLs:
URLURL TypeDescription Paper
Cao, Yulong0000-0003-3007-2550
Xu, Danfei0000-0002-8744-3861
Weng, Xinshuo0000-0002-7894-4381
Anandkumar, Anima0000-0002-6974-6797
Xiao, Chaowei0000-0002-7043-4926
Pavone, Marco0000-0002-0206-4337
Record Number:CaltechAUTHORS:20221221-004623591
Persistent URL:
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:118535
Deposited By: George Porter
Deposited On:22 Dec 2022 18:36
Last Modified:02 Jun 2023 01:28

Repository Staff Only: item control page