"LiDAR Sensors Have Fixable Security Vulnerability"

A large blindspot in front of an approaching autonomous vehicle's LiDAR system can be created by shining expertly timed lasers. This attack can hide moving pedestrians and other obstacles. A group of researchers from the University of Florida, the University of Michigan, and the University of Electro-Communications in Japan discovered this security flaw, which tricks LiDAR sensors into deleting data about pedestrians and obstacles. The team also offered upgrades and fixes to keep LiDAR sensors safe from malicious attacks. According to Yulong Cao, a Ph.D. student at the University of Michigan, LiDAR sensors work by firing laser pulses and calculating the distance from the sensor and potential obstacles in the car trajectory based on the timing of returned signals. However, due to the noisy signals in the real world, LiDAR sensors used in autonomous vehicles typically prioritize the first/strongest signals received and filter out signals returned too soon (i.e., close reflections). By generating laser pulses to imitate fake returned signals closer to the LiDAR sensor, attackers can induce the automatic discard of returned signals from real pedestrians or cars on the road far away. Since the sensor automatically filters out the fake signals, the attack goes unnoticed, and real obstacles are no longer perceived. The LiDAR still receives genuine data from the obstacle, but it is automatically discarded because the sensor only sees the researchers' fake reflections. The group used moving vehicles and robots to demonstrate the attack, with the attacker device located about 15 feet away on the side of the road. However, this can theoretically be accomplished from a greater distance using more sophisticated equipment. The technology involved is simple, but the laser must be perfectly timed to the LiDAR sensor, and moving vehicles must be carefully tracked to keep the laser pointing in the right direction. Updates to LiDAR sensors or software that interprets raw data should address this security flaw, and the researchers recommend that manufacturers begin teaching software to look for the indicative signatures of spoofed reflections added by a laser attack. This article continues to discuss the demonstrated attack that can hide objects and fool autonomous vehicles' obstacle detectors by using the automatic transformation and filtering processes of LiDAR sensor data integrated with autonomous driving frameworks.

Laser Focus World reports "LiDAR Sensors Have Fixable Security Vulnerability"