Cybersecurity in the Autonomous Vehicle Realm

This post is also available in: עברית (Hebrew)

As cyber-physical systems (CPS), such as autonomous vehicles, robotic systems, smart grid and unmanned aerial vehicles are becoming more complex, they also grow increasingly more vulnerable to cyber attacks. In cyber-physical systems, the physical and software components are deeply intertwined, each operating on different spatial and temporal scales, interacting with each other. While cyber attacks mainly concern software, attacks on cyber-physical systems can risk people’s safety.

Researchers from Florida Polytechnic University collaborate with Florida International University researchers in using new types of cyberattacks to improve the security and resiliency of CPS. The team studies this imminent threat by defining CPS’s weaknesses and designing stronger systems against different types of threats.

The research is also part of Florida Polytechnic University’s Advanced Mobility Institute’s (AMI) efforts to develop and test autonomous vehicle technology in the state.

One of the possible attacks is called time delay switched (TDS) and consists of a hacker inserting a time delay into a networked control system. In autonomous vehicles, this could affect the critical immediacy in which controllers receive information, affecting sensor function and potentially leading to accidents.

“These attacks represent major safety issues in the future,” said Dr Arman Sargolzaei, assistant professor of electrical engineering at Florida Polytechnic University. “If a bank account is hacked, that causes serious financial problems. But if an autonomous vehicle is hacked, we could be talking about life-threatening consequences.”

The idea of a TDS attack was invented by Sargolzaei and his collaborators, and there is a patent on ways to detect and respond to it, according to traffictechnologytoday.com.

Another possible attack being used to test the CPS is called false data injection. Part of the project creates a Random Attack Generator that combines TDI with false data injection, to generate many possible types of intrusions to the system. “Hackers might use these new kinds of attacks in the future and we don’t want to wait for that to happen to then figure out what to do,” said Sargolzaei. “We need to be creative on how we can get ahead of them.”