What is the key difference between these two terms? To begin with, you should have a broader look into four words: ‘autonomous’, ‘self-driving’, ‘driverless’, and ‘automated’.
There are different terms for motor vehicles which can operate and control the direction and speed themselves and, maybe surprisingly enough, all of them do not mean the same thing.
‘Driverless’ is likely the most understandable of the terms – it is necessary for a vehicle without a driver to be capable of doing everything by itself in its designated use’s area. That is why the model has to be fully automated. Also, as ‘fully’ hints in the case of automation, there are, believe it or not, different levels.
The different levels of automation: What are they?
The Society of Automotive Engineers (SAE) specifies six levels of automation, starting at no automation (Level 0), and leading up to full automation (Level 5).
You can find the simplest automated tasks in Level 1, with attributes including parking assistance and adaptive cruise control.
Here, while automobiles can fulfill particular duties, there still requires a driver to exercise the ability to direct the vehicle at any time.
Though the system can drive the car by itself on specific roads and under particular conditions, it is still significant for the driver to supervise what the motor vehicle does.
Instead of the original equipment manufacturer (OEM), the driver is responsible for road safety, which makes it far simpler for an OEM to bring the system into play since it needs less validation. With that being said, crucially, the driver understands that she or he has to supervise the automobiles even when it feels as if the car can handle everything.
Here, things get even trickier. When it comes to this level (also known as conditional automation), the vehicle should be able to control its environment entirely, yet can fall back on the driver.
So, though the driver does not have to focus, he or she should be fully prepared to take control at any time. It is a challenging task as the motorist might have looked away from the roadway and is bound to assess the situation if called upon.
What is more, this level is tricky for the system developer, because SAE has not been specific about how much time is necessary for the driver to come back into the loop. Well, do you usually hear statements such as “a few seconds”? In case you missed it, each second at 120km/h is equivalent to 33 meters travelled – that is why any emergency situation will come towards you as the driver exceptionally fast (or vice versa).
In this regard, different OEMs have been making different design options, and it is fascinating to watch what the operating constraints of the systems called ‘Level 3’ are as well as in which cases they require the driver to help.
Things become more explicit, albeit not simpler to implement or validate, in other automation levels.
Level 4 is also known as high automation; the car with this system should be capable of doing everything by itself without the involvement of the driver in particular so-called “driving modes”.
By a driving mode, we mean a traffic scenario – for example, highway driving or maneuvering in the parking garage. It is Level 4 that the automation Google (with Waymo, their self-driving car project) is striving for: cars that will be able to drive on a specific mapped road network in a particular area.
Level 4 vehicles need to have the proven ability to tackle all situations, yet they may, for instance, never enter a highway. Understandably, their sensors are not intended for the specified range to move at high speeds. The more quickly the car goes, the further its sensors need to see to take the automobile to a safe stop if the driving path gets blocked.
“Driving mode” is considered the key to the highly-automated vehicles’ introduction. On the one hand, it is relatively easy to build Level 4 self-parking vehicles for garages, particularly if humans are not enabled to walk in such a car park.
On the other hand, making this work over a small city is much harder. There will be challenges when it comes to releasing these vehicles for regular mobility services.
In the end, there is full automation called Level 5. According to SAE, this system should be effective at dealing with every driving task in every driving mode and under any environmental condition, just similar to a human driver.
While in snow or fog, the system at Level 4 would not work, a vehicle at Level 5 will necessarily master even these adverse conditions from the weather. Lousy weather is known as a primary enemy to car sensors (and to actuators now and then); hence, a lot of vehicle automation experiments do not come out in sunny Nevada, California, or Arizona for no reason.
Sure, Level 5 is the most challenging to attain, yet unfortunately, a lot of business cases for new mobility forms only make sense when vehicles are fully automated. It is difficult to imagine automated Uber fleets becoming grounded only due to snow in New York City.
So, Which Terms Are Appropriate To Which Level of Automation?
You can call a vehicle starting at Level 2 autonomous. It is because it can make its own decisions about driving.
Based on this argument, calling it self-driving is okay as well, although the term tends to be more adequate for automobiles at Level 4 and 5. Not all, these are the only levels that a motor vehicle can get driverless, which conversely is not synonymous that it must not have a driver ever. Similar to the aeroplanes’ autopilot, the car’s autopilot may feature an on/off switch.
Plus, just because a car can carry out the driving task itself is not equal to that the driver expects it to do so. Even with escalating traffic density, you will still see situations in which it is pure fun to drive, and drivers desire to be in charge.
Even though factually different, the two terms ‘autonomous’ and ‘self-driving’ will continue to refer to the same thing culturally. Confusing terminology aside, let’s patiently wait and see how this autonomous technology keeps progressing.