According to the US National Highway Traffic Safety Administration, human error is the factor responsible for 93% of car crashes. Consequently, if human error is eliminated, car crashes will be diminished and our roads will become much safer.
Self-driving cars could be the answer to this major issue because computers can respond much faster to unexpected events that may occur while driving.
So in the near future, we could say goodbye to drunk-driving, playing with the phone while driving, or other dangerous behaviors behind the wheel.
Simply put, technology can save lives.
Not only that, but autonomous vehicles will transform the way we drive and the idea of ownership.
By 2030, we will most likely shift from the car ownership model to mobility as a service.
Imagine this: one morning, you wake up and you need to go to work. You don’t own a car so what you’re going to do is request an autonomous car to come and pick you up.
How crazy does that sound? No driver, no small talk needed. You can even answer a few emails until you get to work.
Or take a quick power nap and not worry about waking up to the driver giving you a strange look.
In this post, I will focus mostly on the technology behind autonomous cars and explain how they are integrated.
SAE International’s On-Road Automated Vehicle Standards Committee described in this report five many levels of automation.
Let’s start with these.
Levels of Automation for On-Road Vehicles
- Level 0. No automation
At this level, the driver performs all the driving tasks. You most likely own this type of car right now.
- Level 1. Driver assistance
The driver still controls the car, but the vehicle may have some driving assist features, such as steering, or acceleration/deceleration. The vehicle, however, won’t be able to perform two tasks at the same time.
- Level 2. Partial automation
In this case, the car can steer and accelerate at the same time, but the driver will still be in control. For example, Tesla’s self-driving system falls into this category.
- Level 3. Conditional automation
A driver is still required in the car, but the vehicle can drive itself in virtually any situation. ‘Virtually’ is a key word here because if the car doesn’t know what to do in a particular scenario, then the driver needs to be ready to take over.
- Level 4. High automation
This level of automation means a driver won’t be required to take over in certain situations or even pay attention to the road. The car will be able to drive and stop in certain conditions (steering, braking, accelerating, responding to events, determining when to change lanes, turn, use signals).
- Level 5. Full automation
The car will be able to drive itself under any conditions, just like a human driver would — which is why this type of vehicle may not even have a gas pedal, brake, or steering wheel.
This is the stage where the driver becomes the passenger.
Chances are that we won’t be seeing fully-automated cars on the roads for ten more years, but they will be here eventually.
Now that we learned about the different types of self-driving cars, let’s find out what kind of technologies these carS use to see everything that’s around them.
4 Technologies Used by Self-Driving Cars
- LIDAR (Light Detection and Ranging)
This technology enables self-driving vehicles to identify the objects that are around them and know the exact distance of these objects in relation to the car.
The sensor is mounted on the roof of the car (you’ve probably already seen it on Uber’s self-driving car).
LIDAR has a 360-degree vision and fires off millions of beams of laser light to measure distances. These sensors are crucial for autonomous vehicles because they can be used to generate 3D maps which will make the environment extremely predictable.
According to John Krafcik, CEO of Waymo Inc, Waymo’s LIDAR technology can go as far as to detect football helmets, aside from pedestrians and cars.
2. Radar
This technology was developed during World War II, and it uses radio waves to measure distances and identify objects that can be as far as 200m away. By tracking an object, this sensor can tell the speed of a car and the direction it’s headed on.
Its performance is not affected by extreme weather conditions, such as snow, fog, or heavy rain.
3. Sonar
Bats use the same technology, if you will, as self-driving cars. They rely on ultrasonic sounds to detect and localize their surroundings.
The sonar (SOund Navigation And Ranging) technology uses sound waves to measure distances and was first developed during World War I. It was used to detect submarines because underwater, the sound travels better than light or radio waves.
This idea was prompted after the Titanic had sunk.
The sonar technology has short-range sensors and it’s mostly used for parking sensors because it can’t measure distances exceeding 60m.
4. Passive optical
This technology uses cameras and intelligent image-recognition algorithms to understand the environment. These sensors can detect colors and the contrast better than the previous technologies I’ve mentioned.
As a downside, it’s worth mentioning that their performance decreases as light diminishes.
LIDAR vs. Passive Optical
Right now, the LIDAR technology, for example, can cost up to 80K per unit. Because of this, companies like Waymo — formerly the Google self-driving car project — have been working on decreasing the cost of the LIDAR sensor.
To get the gist of the costs, Google disclosed that in 2012, their self-driving car used extra equipment that cost $150,000, which was way more than the car itself.
Google aims to reduce the cost to up to 90% in order to scale this technology and make it reliable so that self-driving cars will be accessible to commercial usage.
On the other hand, Elon Musk, Tesla’s CEO, says about the LIDAR technology: “I just don’t think it makes sense in a car context. I think it’s unnecessary.” He believes that for full autonomy, a car would need 360 cameras.
We shall see in a couple of years if Elon Musk was right and the Tesla self-driving cars can be reliable even without LIDAR technology.
Final Thoughts
Audi, BMW, Ford, Nissan, or Toyota are already developing cars that have some autonomous capabilities, but it may take more than five years until fully-autonomous vehicles will become available.
Most of these car manufacturers predict that full self-driving vehicles will be on the road by 2025.
And even after that, the roads will not become packed with self-driving cars overnight because not everyone will rush out and buy a new car.
All assumptions aside, there’s one thing we know for sure: self-driving vehicles will carry enormous consequences for automakers, tech companies, insurance companies and for many other sectors.
Let’s find out together what the future will look like! Follow me on Medium and on Twitter to keep up with the latest news in the industry.
