Resolving Privacy In The Car

As we increasing connect our vehicles — and therefore ourselves — with the outside world while driving, concerns about privacy are coming up, and for good reason.

As I was researching software in the vehicle for this article, Larry Lapides from Imperas Software brought up a good point about how security and privacy in the vehicle will likely be split into two separately addressed issues.

And while he’s not exactly how it will all roll out, he said, “think about the autonomous car driving in a network of autonomous cars…all of these cars are on the road and they need to communicate with each other, they need to communicate with the infrastructure to let someone know to let the other cars know, let the AI know where they are — so there’s got to be security so that can’t be hacked. At the same time you don’t necessarily want someone tracking where you’ve been when you’re out of the house. It may be similar technology that’s going to be used for this or it may just be at the application level.”

Lapides is also self-admittedly old-school about this but said there are privacy concerns, and “we really don’t want the full big brother in the sky and that sort of thing.”

On the technology side, these concerns are being taken seriously, and according to Andrew Patterson, automotive business development manager at Mentor, a Siemens business, there are two ways to deal with privacy and security. “You can either have multiple domains on your device to be super safety critical where security is needed, and put those on a small certified operating system, then run alongside it a more complex ‘normal world’ operating system that is less certified or less tested. That’s been one approach where you’d use a hypervisor or separation or just run two operating systems with two separated applications. But as we get able to certify more complex software systems it becomes more acceptable to run those complex applications. That’s the trend that’s now happening — people are looking at ways to push that complexity barrier up and achieve safety testing and safety certification.”

Of course particularly in the area of autonomous vehicles the safety topic is right at the top of the agenda for all automotive OEMs. “Algorithms that make decisions then have to be acted on in a safe way so if you’re looking at a row of trees versus a queue of people, and you have to crash into one of them, you need to make the right decision. You can never fully safety certify an algorithm but you can definitely safety certify the outcome and what happens as a result of that algorithm,” he added.

While it is encouraging to know these issues are being considered, more discussion needs to happen at every level of the automotive ecosystem. But by the time full autonomous driving arrives, I’m going to trust it’s all been resolved.

Related Stories
Autonomous Cars Drive New Software Requirements
Security, reliability and complex integration are required, but can it happen quickly enough?
LiDAR Completes Sensing Triumvirate
Technology will complement cameras and radar in autonomous vehicles.
What Can Go Wrong In Automotive (Part 2)
Understanding security risks, ECUs vs. SoCs; dealing with an explosion in data.