Developing a Unified Evaluation System for Analyzing User Awareness in Response to Adversarial Image Attacks on Autonomous Vehicles
Anjali Pullareddy
Undergraduate Researcher
Operations Management Major (Kelley School of Business)
Xiaojing Liao
Faculty Mentor
Xiaojing Liao (Luddy School of Informatics, Computing and Engineering)
Project Description
As autonomous vehicles become more prevalent, ensuring their security and resilience against adversarial attacks is crucial. Adversarial image attacks, which involve subtly modifying images to deceive AV perception systems, pose a significant challenge. To address this issue, it is essential to assess user awareness and responses to such attacks. By developing a unified evaluation system, we can gather valuable insights into user perception, behavior, and understanding of adversarial image attacks, ultimately enabling the enhancement of AV security and the development of robust defense mechanisms.
Technology or Computational Component
The computational or technological component of the project involves (1) generating and deploy adversarial images that mimic potential attacks on AV image recognition systems. (2) designing and implementing a human subject study to assess user awareness and responses to adversarial image attacks.