Headline | Autonomous vehicle detection systems can be breached with ‘phantom’ images | |
---|---|---|
Year | 2020 | |
Month | February | |
Country | Israel | |
Description | Ben Nassi, a researcher at Ben-Gurion University of the Negev, and his team demonstrated how easy it is to fool the driver assistance systems that are built into autonomous and connected vehicles. They show that a projector can be used to mislead the system into detecting a projected phantom image as if it were a real signal or a real danger. The systems tested were the Mobileye 630 PRO, used in cars such as the Mazda 3, and Tesla’s HW 2.5 autopilot system. Both have an automation level 2, i.e. they can act as an autopilot but require a human driver for monitoring and intervention. | |
Intentionality | Hacker | |
Target | Vehicle | |
Company | Tesla, Mazda | |
Model | TESLA MODEL X, MAZDA 3 | |
Type of company | OEM | |
Data / Life | Life | |
Access | Remote | |
System | Sensors | |
Recognized by brand | No | |
Source | https://www.nassiben.com/phantoms |