The snow storm on the highway in southern Sweden left thousands of drivers stranded for hours. It was a minor disaster with a lot of suffering and lost resources. Afterwards, it was reported that Halmstad University is researching how AI – artificial intelligence – in self-driving cars could prevent such scenarios in the future.
“Self-driving vehicles drive in California, for example. There are no snowstorms there, but AI is already being tested in cars in places north of the Arctic Circle to develop the technology to work in all weather conditions. Although AI technology has advanced incredibly fast in the last year alone, it is still quite undeveloped and limited.
AI in cars
If you say you have a car with AI, many people will think you have a self-driving car, or at least a car with advanced driver assistance. However, the history of ‘intelligent’ cars with automatic functions goes back at least to the 1950s. That’s when automatic cruise control came along, and it certainly felt very advanced and modern.
The 1970s saw the introduction of electronically controlled fuel injection and anti-lock braking systems (ABS). These things alone helped make driving more comfortable, safer and improved the performance of cars.
In the decades to come, ‘expert systems’ were developed in cars that allowed the car to take in information from sensors, and make logical decisions. This could be used in automatic transmissions and suspension systems.
However, it is perhaps the GPS systems, collision warning systems and parking assistance that you really associate with AI, and with the latest AI technology, this is really getting good.
For example, a number of car parking sensors allow the AI system to provide you with rock-solid and relaxed parking.
Integration of AI in all car brands
Today, most car brands have advanced driver assistance features. In fact, all car brands will probably have to keep up in order to compete.
After all, an ‘intelligent’ car like this needs to be able to see, hear and feel for itself, so AI cars are equipped with a variety of sensors and cameras, among other things. Cheap technology for sensors, for example, has made this possible.
Tesla is a pioneer in AI and back in 2012 it offered cars with a system called Autopilot 1.0 that could, for example, suggest lane changes. Since then, it has continued to take the lead, and in 2019, cars with its new system, FSD, were introduced, enabling automatic city driving, for example.
Today, however, many other car brands have caught up, with BMW, Mercedes, Ford, Chevrolet and many other brands offering cars with active lane keeping assistance, automatic emergency braking, voice control and facial recognition. In fact, the most advanced systems allow drivers to drive without even holding the steering wheel!
Easier parking, relaxed long-distance travel, safer driving.
So why is this technology so attractive to drivers? It’s about both convenience and safety.
Parking was mentioned earlier, and being able to pocket park or park in tight parking lots without hardly thinking probably sounds appealing to most drivers.
Long-distance travel may be fun and exciting, but it’s also tiring. Imagine being able to literally put the car on autopilot, and the car steers itself, changes lanes and overtakes.
Now, as I said, these self-driving features are limited, but they are rapidly becoming more common. There are more and more reports of people finding (illegal) ways to drive the cars without even being in the driver’s seat!
AI for safety
Overall, this technology contributes a lot to safety. After all, ABS brakes and collision warning systems have helped drivers avoid accidents for many years. There is now often a complex trade-off between how much cars should be allowed to act on their own and when and how the driver must take over.
Part of the research now is about how to make these cars ‘smarter’ and actually make decisions that are equivalent to a good driver’s and even better. In urban traffic, for example, a cyclist can be temporarily hidden by other vehicles, and even the best self-driving cars interpret that as the object not existing. A human driver continues to take the cyclist into account.
Research is ongoing in this area, and the question of how AI will actually lead to much safer traffic is a high priority. As mentioned in the introduction, it is hoped that AI in cars will be able to avoid traffic jams in snowstorms and the like in the future.
This and many other things will certainly be handled by future cars with AI.
