From Tesla to Xiaomi, the future is driverless – but how do autonomous cars actually work?

The race to full autonomy is in full swing, with new brands launching self-driving cars every day.

Different brands are using different ways to get to the same result, but there’s no clear winner yet.

However, there are at least two ways we can break down self-driving cars.

We can break down the technology and we can also break them down into different categories.

Level 0 to Level 5, this is how self-driving cars are defined

Conventionally, automakers agree that there are six levels of autonomy, ranging from Level 0 to 5.

With Level 0, the car has no autonomy at all, with no driver-assistance systems.

Then we get to Level 1 and Level 2.

Both require human control but both use some form of driver assistance, which is very basic in Level 1, for example cruise control, but gets better and more sophisticated with Level 2, perhaps with the addition of things like lane control and obstacle detection.

With Level 3, it’s the machine, not the driver, that controls the driving environment, and the car can perform most, this is the key word, driving tasks.

With Level 4, the vehicle can do everything by itself, it can perform all tasks, and the vehicle can theoretically go from A to B by itself.

The key thing that Level 3 and 4 have in common is the fact that, in both cases, a human driver that’s ready to override is required.

With Level 5, the vehicle is fully autonomous, with no ifs or buts or clauses.

You can get rid of the steering wheel, get rid of the pedals, and the passenger can fall asleep after getting into the car and wake up when they arrive at destination.

For reference, Mercedes is currently operating at Level 3, while Waymo and Tesla are operating at Level 4.

Ideally, at least in theory, Cybercab can reach Level 5 at some point.

Lidars vs cameras

There’s an interesting debate when it comes to self-driving cars.

We’ve got two factions, so to speak, with one advocating for lidars and while the other wants to use nothing but cameras.

Most automakers are using a combination of cameras and lidars, which are essentially improved radars with things like heat and light detection.

By contrast, other companies like Xpeng and Tesla are using cameras but no lidars.

This is because Tesla and Xpeng argue that lidars have technical limits, while cameras, especially with AI, don’t.

Translated into plain English, using a lidar, according to them, is the same thing as blindly following the sat nav instead of simply looking at the road ahead.

If you’ve got good cameras, enhanced by AI, you don’t need a lidar to tell the difference between, say, a deer or a lamppost.

No one knows who’s right or wrong, yet, but we’ll find out soon enough.

DISCOVER SBX CARS: The global premium car auction platform powered by Supercar Blondie