For self-driving vehicles to become a reality, they need to be connected in order to communicate with their surroundings and with each other. However, this interconnection must be designed in way that protects the vehicles from cyberattacks and protects the privacy of vehicle owners. This necessary to ensure public confidence in and acceptance of self-driving vehicles. It places demands on automotive manufacturers to complement their proficiency in safety with the equivalent in cybersecurity.
A self-driving vehicle that functions as it should is likely to be significantly safer than vehicles driven conventionally. But what happens if the vehicle is affected by external signals that make it think the road is two metres to the right? Of if it thinks it is on a three-lane motorway instead of amidst urban traffic?
– “These types of problem are referred to as spoofing attacks,” explains Tomas Bodeklint, research and business developer at RISE. “This means that someone on the outside broadcasts incorrect signals in order to take control of or otherwise affect the vehicle. It can have disastrous consequences and seriously harm public confidence in self-driving vehicles.”
Automotive industry needs cybersecurity expertise
Automotive manufacturers are generally highly adept at vehicle safety, and have developed technologies such as advanced driver-assistance systems (ADAS). Jonny Vinter, director of reliable systems at RISE, believes that when it comes to protecting vehicles against external attacks, automotive manufacturers still have a way to go:
– “The automotive industry obviously recognises the importance of this, but the problem is complex and it is not completely understood how to ensure that vehicles are and remain cyber secure. The industry needs to bolster expertise in cybersecurity through research into the area, and by training personnel in both cybersecurity and vehicle safety.”
In the EU-funded SECREDAS project, industry and researchers from around Europe are working to, among other things, increase public confidence in connected and self-driving vehicles. For RISE, much of the work involves testing systems and software and proposing practicable measures.
– “We work a great deal with fault injection, simulating various fault models, and bombarding electronics and software with these fault models in order to evaluate error handling mechanisms,” says Vinter. “But we are also developing and improving fault injection methods to simulate cyberattacks and the measures that will handle them. The goal is a fault-tolerant, cyber-secure, and resilient system able to handle attacks and faults in electronic systems in order to avoid critical situations in which people may come to harm.”
Globally unique test facility
These sorts of tests require a special environment in which to be carried out safely. In Borås there is Awitar, a globally unique test facility that is fully shielded from electromagnetic radiation. The shielding means that, on the one hand, no signals can penetrate the area, but – above all – signals cannot get out.
– “In Awitar we can transmit any signals we wish, without being concerned about disrupting external systems,” says Bodeklint. “This, combined with the fact that the chamber is large enough to accommodate a truck or bus, allows us to test the capability of essentially any connected vehicle to repel external signals.”
Among other things, the chamber enables GPS or GNSS positioning signals to be affected, which can cause a vehicle to think it is in a completely different location than Borås.
– “We can place a vehicle on rolling tracks in Awitar and then drive it as if it were in Stockholm or Singapore. The possibility of testing in a controlled environment how connected vehicles behave when affected externally means we can help the automotive industry to develop vehicles that are both safe and cyber secure,” concludes Bodeklint.