Abstract

ABI Research predicts that 60% of new cars will be internet connected by 2017. When mobile devices communicate with cars, the connected auto becomes a new threat target. Apple's CarPlay and Google's Android Auto interfaces will also bring more integrated, but potentially vulnerable mobile apps into the vehicle. Unfortunately, security vulnerabilities existing inside both the auto's CAN BUS and within the mobile apps can introduce new security risks ranging from unauthorized data capture, to more serious offenses such as vehicle or property theft, to criminally malicious hijacking or even the possibility of remotely overriding critical auto systems and control, resulting in accident, injury, or even death. Various attack examples have been demonstrated at several security industry conferences.
Auto malware is coming at a fast pace. Malicious auto mobile apps can send potentially dangerous CAN message commands into cars via the onboard OBD-II port to control or change the auto's systems and/or status. Drivers may be at risk of being targeted by auto hackers, such as thieves, phishing scams, or nefarious phone attacks. There is currently no solution on the market to specifically defend against such attacks and monitor the communication channels between cars and mobile apps.
Auto malware don't need to damage the car, instead they can leverage traditional phishing or spam tricks to fooll drivers, filing the gap between traditional mobile malware and auto domain. In this abstract, we successfully implemented several auto malware demo. For instance, we implemented the first mobile app of auto over-the-air attack. Another example is able to brute-force scan surrounding OBD dongle, measure the distance between mobile and dongle. When the driver is going to leave the car, the malicious code will send "door-open" or "trunk-open" command to car without driver's awareness. When the door is open, the code will send message to remotely hacker about the car's location so that the hacker will come to steal belongings in the car. In this way, we can build up the auto attack scenario database, and use the scenario database for auto security penetration testing.
How to defend against such attacks? We showcase our research findings on the first auto anti-hack solution, adding protection where there presently is none today on the vehicle system to minimize penetration from outside cyber attacks. Our auto security solution can discover and prevent a range of such attack scenarios in real time, including the attacks mentioned by Charlie Miller at Blackhat and Defcon. Our solution is generic, which means firewall signature can be updated from the cloud for different car models.
How is our research work different with that from other researchers?
Their solution: . only passively monitor data traffic and are incapable of blocking attacks from the unprotected CAN BUS . anomaly detection method will result in high false positive alerts . often incompatible with current commercial car diagnostic devices Our differentiators: . the first auto anti-hacking solution from security and auto industry veterans . inline CAN BUS protection with updateable threat signature database . low false positive . prevention of Denial of Service attacks . Generic protection, but also customizable for specific car models