Engineering researchers at UW look at safety and security of automated vehicles

·2 min read
UW's Autonomous Vehicle Research and Intelligence Lab is teaming up with vehicle supplier Magna International for a project that will look into the security and safety of automated vehicles.  (John Locher/Associated Press - image credit)
UW's Autonomous Vehicle Research and Intelligence Lab is teaming up with vehicle supplier Magna International for a project that will look into the security and safety of automated vehicles. (John Locher/Associated Press - image credit)

University of Waterloo is taking a step closer in the future development of automated vehicles by partnering with vehicle supplier Magna International.

As vehicles become increasingly more advanced and autonomous, researchers with the school's Autonomous Vehicle Research and Intelligence Lab will be looking at how to make automated vehicles safer and more secure.

Magna International will be providing a vehicle for the project, which Sebastian Fischmeister, a professor in the Department of Electrical and Computer Engineering, said will be a major benefit to their research.

"Usually cars are closed systems. As a researcher you can't just go into a car and replace something and try out your own thing," Fischmeister said.

"The messages that are transmitted in the car are not public, you don't know what the bits and pieces mean."

Challenges with automated systems

Automated vehicles are different from fully autonomous vehicles, where the driver can sit back and let the vehicle drive itself, Fischmeister explained.

An automated approach is more incremental — the vehicle automates one driver function after another until the vehicle reaches a point where the driver doesn't have to pay attention, he said, which can present its own set of challenges.

"When you use any aid function like for example, your adaptive cruise control or lane centring, you have the expectation that the car doesn't crash, that the car doesn't do any bad action," he said.

"What we specifically look into is how can we use modern software and system paradigms like for example, machine learning, in a context that you can trust the system."

And an added challenge to that, Fischmeister said, is making sure those systems remain secure.

"Now that all these vehicles are connected, you have the risk of someone sitting somewhere and through the internet and through connectivity reaching out to all vehicles of a particular manufacture," he said.

"That's why safety and security are the key challenges."

Fischmeister adds that by having access to the vehicle, Magna's expertise and knowledge of automotive systems, it means researchers can now conduct their own experiments to improve the system and test how they will perform in real life.

"We can, for example, go in and replace the camera with a more experimental system and try out and see how well our research framework and research results apply in this particular case," he said.

You can hear the full interview below:

Our goal is to create a safe and engaging place for users to connect over interests and passions. In order to improve our community experience, we are temporarily suspending article commenting