Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!
Explore a cutting-edge research presentation from USENIX Security '23 that delves into Physical Removal Attacks on LiDAR-based autonomous vehicle driving frameworks. Learn about a novel method that exploits laser-based spoofing techniques to selectively remove LiDAR point cloud data of genuine obstacles at the sensor level, causing autonomous driving obstacle detectors to fail in identifying and locating obstacles. Discover how this attack, invisible to the human eye, can deceive autonomous vehicles' perception systems by leveraging inherent automatic transformation and filtering processes of LiDAR sensor data. Examine the effectiveness of these attacks against popular AV obstacle detectors, fusion models, and their impact on driving decisions using industry-grade simulators. Gain insights into the limitations of existing defenses and potential strategies to mitigate such attacks, crucial for enhancing the security of autonomous driving systems.