What does 2022 have in common with the 1939 World’s Fair? It’s the feeling that autonomous vehicles are right around the corner. Nearly every automaker, from Tesla to Ford, has overpromised and underdelivered on their plans for autonomous driving. Tesla CEO Elon Musk said it himself recently in an interview with researcher Lex Fridman. “I thought the self-driving problem would be hard, but it’s harder than I thought. I thought it would be very hard, but it was even harder than that,” Musk reflected. He’s not the only one to misjudge the enormity of the task at hand. You’d think automakers would stop giving themselves deadlines that are destined for letdowns.
It turns out that there’s been one overarching theme in the learnings from the past decade of development. Engineers now see that in order for self-driving cars to be safe and successful, cars will have to learn to think like a human. Computers are exceedingly good at performing repetitive tasks. What they’re not great at is responding to unique situations full of unknowns. The human brain is more capable than some give it credit for. We’re very good at dealing with unknowns and making complex decisions on the fly.
After reading this article, you’ll better understand what self driving cars are, the difference between self-driving cars and autonomous vehicles, and the terms and jargon associated with self-driving cars. Let’s dive in.
What Is a Self-Driving Car?
As you’re soon about to learn, the world of autonomous cars and self-driving technology is full of terms worth defining. For starters, what is self-driving in the world of transportation? Self-driving cars can drive in some or even all situations without driver input, but a human must always be ready to take control. Think of them as a crucial stepping stone on the path to full autonomy.
Self-driving cars are not fully autonomous. In the world of professional engineering, ‘automation’ is the preferred term for the sliding scale of vehicle operation status. In fact, the United States National Highway Traffic Safety Administration (NHTSA) refrains from using the term “self-driving” at all, and most engineers disagree with the use of the term. Let’s dive into the terms and definitions that relate to so-called self-driving cars, and the future of automation as a whole.
Terms and Definitions: A Self-Driving Primer
First, let’s clear the confusion. Talking about automated cars warrants a glossary of its own. Being well-informed is the key to knowledge, and we all know knowledge is power. Here’s a list of the terminology you’re likely to encounter in a self-driving world. In most cases, the difference is in the finest of details.
Advanced Driver Assistance System (ADAS)
These are technological features designed to improve driving safety. These software-based systems improve a driver’s ability to react to adverse situations on the road. Examples include adaptive cruise control, forward collision warning and lane departure warning. ADAS features are very common in newer car models.
Autonomous or Automated Vehicle
This term is thrown around a lot, but a true autonomous vehicle is capable of sensing its environment and operating without human involvement. Human passengers can take their eyes off the road and just enjoy the ride. Imagine entering the destination address, and that’s it. Autonomous vehicles can do everything that an attentive human driver can do.
Usually referring to Tesla Autopilot, which is a suite of ADAS features that enable the vehicle to steer within a lane and adjust speed in response to surrounding traffic. It’s essentially adaptive cruise control plus lane centering. Autopilot is standard on all new Tesla models.
Tesla “Full Self-Driving” in its current iteration is not much different from driver assistance technologies. Tesla enthusiasts, relax. That’s likely to change as Tesla updates the software regularly via over-the-air updates that simply require WiFi to install. Automotive engineers generally refrain from using this term altogether, as it’s more associated with marketing than with engineering automation. A true full self-driving car can navigate roads with human supervision. Tesla’s program is getting close, but as thousands of videos online will show you, Tesla FSD is not quite there yet. It is impressive though. Tesla FSD can be yours (someday) for an additional $12,000 when you buy a new Tesla vehicle.
This is when a vehicle’s operation is limited to a restricted geographic area. For example, Waymo’s driverless vehicles are geofenced to only operate in the Phoenix metropolitan area. Geofenced autonomous vehicles will grow in popularity before true, independent autonomous vehicles put rubber on public roads.
Light Detection and Ranging. LiDAR is the go-to ‘radar’ technology for nearly all self-driving innovators. Except of course for Tesla, who seems to think image processing with cameras is the way to go. LiDAR can ‘see’ through low visibility conditions, including fog and heavy rain.
When a vehicle can perform most driving tasks in a geofenced area, but with constant human monitoring and intervention when needed, some in the automotive industry consider it to be self-driving. The degree of human input varies. More on that below…
The Six Levels of Automation
But wait, there’s more! Engineers rate the levels of automation based on how independently the system can perform tasks, and how much human input and supervision is required. Here are the basics of the five levels of automation, according to the Society of Automotive Engineers.
- Level Zero – Limited to warnings and brief takeover of vehicle control. Ex: auto emergency braking, blind spot warning, lane departure warning
- Level One – Steering OR acceleration/braking. Ex: lane centering OR adaptive cruise control
- Level Two – Provides both steering AND acceleration/braking. Ex: lane centering AND adaptive cruise control
- Level Three – The car can drive independently under certain conditions. Ex: automated driving at slow speeds
- Level Four – Geofenced automated driving; steering wheel optional. Ex: Waymo’s local driverless taxi
- Level Five – The autonomous vehicle can operate anywhere without driver input or attention
Now that we’ve covered the engineering and industry jargon, let’s revisit our definition.
What exactly is self-driving technology? The most agreeable definition is that self-driving cars fall within level 3 or level 4 automation, in which the vehicle can perform most driving tasks in a geofenced area, but with constant human monitoring and intervention when needed. Fully autonomous vehicles fall within the ultimate frontier of automotive engineering, Level 5. A true autonomous vehicle can operate from start to finish without driver input or attention. Imagine reading a book or taking a nap on your way to work.
Are There Any Self-Driving Cars Today?
No cars on today’s roads are capable of fully autonomous driving. Automated driving remains years away, however tremendous resources are committed to the cause of unraveling the ultimate challenge in automotive engineering. Tesla’s Autopilot and Full Self-Driving features are branded as Level 2 systems, which means that constant supervision is required, and intervention is to be expected. This discrepancy between the Level 2 classification of Tesla’s driver assistance systems and the names of the products remains the source of much controversy among engineers and driver safety advocates alike. Can’t we all agree that honest advertising is always in the interest of safety and responsibility?
Aside from automakers, there are dozens of other companies innovating in the autonomous driving space. Waymo, Argo AI and Cruise are all putting geofenced autonomous cars on the road today for real-world testing and limited customer use for ride-hailing. What is a self-driving car in 2022? It’s likely a prototype with limited use.
Tesla is the clear leader in advanced driver assistance systems. However, the extent of Tesla’s lead among industry competitors is not nearly as clear as it was a few years ago. Tesla has taken a bold step away from using radar for sensory inputs. The decision to remove LiDAR from new Tesla models starting in 2020 was so controversial that some senior engineers quit in protest.
Remarkably, Tesla is no longer the only automaker breaking autonomy barriers. In 2021, Mercedes-Benz became the very first automaker to get regulatory approval for Level 3 autonomous driving on limited public roads. For now, Mercedes Drive Pilot is available on 8,197 miles of German highways at speeds up to 37 miles per hour. What makes Mercedes Drive Pilot so special is that it is the first approved consumer-ready system to permit the driver to take their attention away from the road while the vehicle is in motion. Even Tesla’s Full Self-Driving feature does not permit the driver to direct their attention elsewhere, despite evidence of the contrary on social media.
Level 2 For Now
For the foreseeable future, American roads will see even more Level 2 driver assistance systems calculating their way through traffic as nearly every automaker in the market steps up their autonomy game. Level 3 remains in development, even for Tesla. Mercedes has not announced if it will seek approval in the US anytime soon, likely due to the murky regulatory environment.
Fully automated driving is likely in our future, but no one knows when it will be safe and accessible to all. The pace of innovation ebbs and flows. Engineers, regulatory agencies and insurance companies have some hard problems to solve. For now, proceed with caution when at the wheel of a “self-driving” vehicle. Their arrival is a great reason to look twice when crossing the street. You never know who (or what) might be coming around the corner.
CarEdge Just Launched a New Search Engine to Make Buying Your Next Car Easier!
We have thousands of vehicles listed today. Each listing includes industry insights, empowering data and the true TotalPrice that will make buying a car the transparent process it should’ve always been. We show you what dealers don’t want you to know. We believe that data can be your superpower when buying a car. Check it out here!