Qihoo 360 Vulcan team receives an award from Microsoft Edge bug exploiting and pwning during the POC 2016 held on Nov 10, 2016 in Seoul, South Korea. (Photo provided to chinadaily.com.cn)
White hats from Qihoo 360 Technology Co have showcased the latest bug-fixing improvements at an international hackers conference.
Organized by South Korean hackers and security experts, the Power of Community (POC) was first held in 2006. The conference, which started on Thursday, has become one of the oldest conferences that specializes in security and hacking with a global perspective.
Both malicious hackers known as "black hats" and ethical hackers known as "white hats" are encouraged to participate in sessions to share and discuss the latest hacking and internet security topics.
Yan Ruimin, a security researcher at Qihoo 360's vehicle cyber security lab, dubbed "SkyGo Team", and Yan Chen, a PhD student at the Ubiquitous System Security Laboratory of Zhejiang University, gave a joint presentation: "Can you trust autonomous vehicles? Contactless attacks against sensors of self-driving vehicles."
They introduced the latest development in Advanced Driver Assistance Systems (ADAS) which can sense the driving environment and warn drivers of immediate dangerous situations. They also discussed Tesla Motors' driverless technology.
They explained how sensors provide a vehicle with measurements of its surroundings and how sensory information is used for navigation.
"Although promising and proven safety features, the trustworthiness of the existing driverless cars has to be examined before they can be widely adopted on the road," Yan Chen said.
They investigated sensors whose measurements are used to guide driving, such as millimeter-wave radars, ultrasonic sensors and forward-looking cameras.
Video clips were collected during the contactless attacks on a few popular autonomous cars, including Tesla, Audi and Volkswagon in both the lab and outdoors.
Using off-the-shelf hardware attackers were able to perform jamming and spoofing attacks, which Tesla to become 'blind' and malfunction. This could potentially lead to crashes and greatly impair the safety of self-driving cars.
They also proposed software and hardware countermeasures that will improve sensor resilience against these attacks.
