Imagine getting into your car, telling it where you want to go and then just sitting back and enjoying the scenery while your car does all the driving. The self-driving car was a staple of science fiction, but technology has now advanced to the point where autonomous vehicles aren’t too far removed from reality. While you may have seen various prototype tests of this technology from one or more carmakers, you may be a little uncertain as to what it all means. Let’s learn a little more about car automation.
The six levels of automation
Automation as it applies to cars is broken down into six levels (0-5) by the Society of Automotive Engineers (SAE). These levels have also been adopted by the U.S. Department of Transportation. Currently, only Levels 0-2 are available for consumer purchase within the United States.
No automation: The human driver is fully responsible for all driving tasks such as steering, acceleration, braking and more. The vehicle may provide some automated warnings, such as a blind spot or lane departure warning but it’s still up to the driver to manually correct for those.
Driver assistance: This is typically a single automated system that provides some continuous support with steering or speed monitoring, such as adaptive cruise control or lane centering. The driver is still required for most driving functions.
Partial automation: Both steering and acceleration can be simultaneously automated. That means you can activate lane centering and cruise control at the same time, unlike Level 1 where it’s one or the other. The driver still needs to actively monitor all tasks and can assume control at any time.
Conditional automation: The vehicle additionally features environmental detection capabilities, allowing them to potentially detect a slow-moving vehicle, for instance, and independently make the decision to overtake it. The vehicle can perform most driver tasks but will require manual override in less optimal driving situations. At the time of writing, this level of automation is currently unavailable for consumer purchase in the United States.
High automation: The car can perform all driving tasks in a broader range of conditions, and human override is rarely required. The vehicle’s operation is generally restricted within a certain area using a practice known as “geofencing.” This level of automation is currently unavailable for consumer purchase.
Full automation: The vehicle requires no intervention from the human driver and can handle virtually all driving tasks in virtually all conditions, with no geofencing restrictions. This technology is currently unavailable for consumer purchase.
How does an autonomous car work?
Now that you understand the basics of car automation, let’s take a closer look at some of the technology that goes into making self-driving cars possible.
Radar sensors around the vehicle help detect objects and their relative distance to the car.
Lidar is another object detection system that works on the same principle as radar but uses lasers instead of radio signals. It’s capable of much finer detection than radar and can also 3D-map objects.
Cameras have been a helpful addition even for “traditional” cars, and they definitely go a long way toward helping self-driving vehicles see things like traffic lights, lanes, road signs and more.
Ultrasonic sensors placed in the car’s wheels can detect curbs and other vehicles, which come in handy when trying to park.
Artificial intelligence (AI)
Sophisticated AI software processes all the incoming data from the sensors above to make decisions about things like steering, acceleration and braking. Hard-coded rules and algorithms allow it to follow traffic rules and avoid objects, other cars and pedestrians.
Pros and cons of self-driving cars
Autonomous cars represent a significant shift in the future of the automotive industry. Change can be good or bad, and self-driving cars are no exception.
Pros of self-driving cars
- Autonomous cars are more likely to follow traffic rules, as this behavior will be built into their programming.
- Machines aren’t susceptible to human error. They can’t misjudge the distance to something, get distracted, fall sick or be too impaired to drive.
- The two above benefits would likely create a third: Ostensibly fewer accidents and a smoother flow of traffic.
- Self-driving technology allows greater mobility for people who may not be able to drive on their own.
Cons of self-driving cars
- Humans use a range of non-verbal cues like facial expressions to inform decisions made while driving. Take for instance, the blind enthusiasm of a young child whose eyes see only the ball they’ve sent bouncing into the road. A human driver might recognize that look and slow down, but would an autonomous car make the same split-second judgement?
- Regulatory processes around autonomous vehicles are still in the works. As such, the legal landscape could change drastically from one moment to another and particularly from one state to another.
- On a related note, insurance regulations around autonomous cars are also currently not fully defined. In the event of an accident, who is liable for what? A Level 5 car, for example, may not even have any manual controls for human passengers to use, which complicates the issue significantly.
- Whenever discussing anything that runs on so much software, there's always the possibility it may become vulnerable to cyberattacks. While this issue is being actively addressed by many carmakers, it remains a tangible concern.
A self-driving car used to be a sci-fi fantasy, but these days it looks like it’s just around the corner. As technologies like radar, lidar, and AI become more common, it’s only a matter of time before you see autonomous vehicles rolling through your neighborhood. The question today isn’t so much whether we can build a car that requires no driver — we can. Now it's a matter of deploying this technology in a way that ensures maximum safety and convenience for drivers, passengers and pedestrians alike.