Ce Zhou, Khang Nguyen, Liyang Xiao, and Qiben Yan, Michigan State University
To achieve a more accurate and robust understanding of the driving environment, researchers have extensively explored sensor fusion algorithms that integrate data from multiple modalities. Among these, the fusion of radar and camera has gained attention due to their low cost and widespread adoption. However, most existing methods are evaluated only under benign driving conditions, leaving their performance under adversarial scenarios largely unexplored. Thus, in this poster, we evaluate five camera-radar sensor fusion algorithms against various black-box adversarial attacks, including Gaussian blur and motion blur induced by disturbances to the camera stabilizer. Our preliminary results indicate that the performance of these fusion algorithms consistently degrades as the intensity of adversarial perturbations increases.
Open Access Media
USENIX is committed to Open Access to the research presented at our events. Papers and proceedings are freely available to everyone once the event begins. Any video, audio, and/or slides that are posted after the event are also free and open to everyone. Support USENIX and our commitment to Open Access.
