Project Description:

Autonomous vehicles are visioned as a revolutionary power for future transportation. It is predicted that by 2045 more than half of the new vehicles manufactured will be autonomous. A fundamental integral component of autonomous vehicles is the perception system consisting of various sensors (e.g., camera, LiDAR, and radar) whose task is to perceive the surrounding driving environments. The state-of-the-art perception systems in autonomous driving mainly rely on Artificial Intelligence (AI) techniques to process complex sensory data and make driving decisions. However, existing AI techniques have been proven vulnerable to adversarial attacks. Many AI models can be easily fooled by the attackers who are able to manipulate the models' inputs. Despite the growing security concern, the study on adversarial attacks against the perception systems in autonomous driving is limited. This project investigates adversarial machine learning challenges faced by AI enabled perception systems in autonomous driving with the aim of formulating defense strategies. In this project, we will develop practical adversarial attack methods that can be used to analyze the security vulnerability of the perception systems in autonomous driving. We will also explore possible defense strategies that can effectively reduce the vulnerability of these systems and make them more robust to adversarial attacks.

Start Date:

July 2023

Postdoc Qualifications:

The candidates need to be knowledgeable in interdisciplinary areas including autonomous vehicles, AI and machine learning, as well as security.

Co-Advisors:

Jing Gao, School of Electrical and Computer Engineering, Purdue University, Email: jinggao@purdue.edu, Web: https://engineering.purdue.edu/~jinggao/

Yiheng Feng, School of Civil Engineering, , Purdue University, Email: feng333@purdue.edu, Web: https://engineering.purdue.edu/CE/People/ptProfile?resource_id=244000

Bibliography:

[1] Adversarial Sensor Attack on LiDAR-based Perception in Autonomous Driving, Yulong Cao, Chaowei Xiao, Benjamin Cyr, Yimeng Zhou, Won Park, Sara Rampazzi, Qi Alfred Chen, Kevin Fu, and Z. Morley Mao, in Proceedings of the 26th ACM Conference on Computer and Communications Security (CCS'19), London, UK, Nov. 2019.

[2] Towards Robust LiDAR-based Perception in Autonomous Driving: General Black-box Adversarial Sensor Attack and Countermeasures, Jiachen Sun, Yulong Cao, Qi Alfred Chen, and Z. Morley Mao, USENIX Security Symposium 2020.

[3] Yi Zhu, Chenglin Miao, Tianhang Zheng, Foad Hajiaghajani, Lu Su, and Chunming Qiao, "Can We Use Arbitrary Objects to Attack LiDAR Perception in Autonomous Driving?", the 28th ACM Conference on Computer and Communications Security (CCS 2021), Virtual Conference, November 2021.

[4] Yi Zhu, Chenglin Miao, Foad Hajiaghajani, Mengdi Huai, Lu Su, and Chunming Qiao, "Adversarial Attacks against LiDAR Semantic Segmentation in Autonomous Driving", the 19th ACM Conference on Embedded Networked Sensor Systems (SenSys 2021), Virtual Conference, November 2021.

[5] Zhi Sun, Sarankumar Balakrishnan, Lu Su, Arupjyoti Bhuyan, Pu Wang, Chunming Qiao, "Who Is in Control? Practical Physical Layer Attack and Defense for mmWave-Based Sensing in Autonomous Vehicles," IEEE Transactions on Information Forensics and Security (TIFS), Vol.16, 2021.