Fully driverless vehicles will be incredibly safe, but we’re not there yet

For decades, science fiction movies and TV shows have been predicting spectacular technological innovations like self-driving cars (as well as flying cars and nuclear fusion vehicles). Every day we are getting closer to cars that can drive themselves and require little to no input or attention from the driver; flying cars are still a long way off (which is probably a good thing, but I’m still bitter about it). You can already find the early stages of self-driving technology in cars, trucks, and SUVs sold right now under impressive names like Autopilot and Supercruise.

Most experts agree that when automation is fully developed, provided the technology works properly, the frequency of failures will decrease dramatically. Even better, when collisions do occur, they should be less disruptive and damaging, with increased safety for occupants of any vehicles. Unfortunately, this will only happen at some point in the future. Right now we have incomplete technology that seems like it can do more than it actually can. Early statistics are starting to show that semi-autonomous driving technology isn’t as safe as we’d like, but that’s not entirely the car companies’ fault.

Standards for self-driving technology

One of the biggest problems with the concept of self-driving technology, at least for now, is that the general public largely misunderstands where we are. To fully appreciate this, it’s important to understand how self-driving technology has emerged and evolved over the past decade or so. The National Highway Traffic Safety Administration (NHTSA) has divided the automation of vehicles that can drive themselves to some degree into six different levels:

• Level 0: Instant assistance to the driver – you are constantly driving and monitoring the car; the car can help briefly with things like power steering.
• Level 1: Assistance to the driver – you drive the car all the time and monitor it; the car provides constant assistance with improved steering or braking.
• Level 2: Additional help – you drive the car all the time and monitor it; the car has a system that can be activated to temporarily handle the steering and acceleration/braking for you.
• Level 3: Conditional automation – the system drives the car and you control the vehicle with the ability to take control at any time; the system handles all aspects of driving, including steering and acceleration/braking, while it is activated.
• Level 4: High automation – the system controls the vehicle, and you, in fact, are the passenger; the vehicle handles all aspects of driving itself, although its use is restricted to certain service locations.
• Level 5: Full automation – the system controls the vehicle, and you, in fact, are the passenger; the vehicle handles all aspects of driving independently in all locations.

When you see a vehicle with a system called “Autopilot”, you might think that we are somewhere at level 3 or even 4. Although such technology is currently being developed and tested in very limited and harsh conditions, we are not there yet . Level 5 is clearly the purpose and ultimate goal of this technology, but in reality we are only at Level 2!

Where we are now

That’s right, despite what much of the automotive market would like to suggest, we’re only about halfway to full automation in vehicles. At the moment, the best technology you can buy in a car falls under Level 2: Additional Assistance. There are no vehicles available to the public that allow you to activate any self-driving feature so that you can take your mind off the road and feel like a passenger when you’re in the driver’s seat. It just isn’t there yet.

The problem, many experts say, is that car companies are pushing advertising and using terms that suggest we’re further along than we are. Names like “Autopilot” make drivers think that the vehicle can take over the driving, but in reality it just provides an enhanced level of assistance while they’re on the road. Vehicles that reach Level 3 and even Level 4 automation are being tested, but none are in mass transit or available to the public. But we are getting there, and when we do, the roads should be much safer.

Semi-autonomous vehicle safety

However, the roads aren’t necessarily that safe at the moment, and all the advanced safety features, including the currently available semi-autonomous driving technology (Level 2), don’t help as much as they should. For example, NHTSA has finally begun to disclose information about collisions and traffic safety with advanced systems and features such as semi-autonomous driving. Last year, they issued an executive order requiring all US automakers to provide crash reports and data to NHTSA after any collision in which advanced safety features were active within 30 seconds of the crash.

This allowed for a much more comprehensive view of how the advanced safety features work on the vehicles and whether they really help. So far, the data is not optimistic and includes the discovery that Tesla cars with autopilot software was involved in 273 crashes reported to NHTSA last year, nearly 70% of the 392 crashes they reported to them during that time. Before the order to provide the information they needed, NHTSA could only investigate 42 crashes that likely involved the use of driver assistance technology over the past eight years. This new requirement has made it much easier for safety organizations to view real-world data and provide it to drivers.

In fairness to Tesla, there were hundreds of thousands of accidents on America’s roads last year, and if less than 300 of them involved Tesla’s Autopilot technology, that’s not bad at all. The biggest problem, according to many experts, is that drivers are still a factor in road safety. Fantastic technology that can improve braking performance, monitor drivers in your blind spot and warn you when someone slams on the brakes in front of you is pointless if the guy behind you is looking at his phone, doesn’t see a stop light and crashes into the back of your car. Also, a shocking number of people pay big bucks for impressive cars with advanced safety features and then turn them off.

What does the future hold?

As long as such additional safety features are not mandatory and are usually only provided on cars that cost more in higher trim levels, things will remain dangerous on the road. Seat belts only started to save a huge number of lives when they became mandatory in all vehicles; similarly, rearview monitors only became a major convenience when every manufacturer had to include them. Automated security features and even full self-management or 5 level of automationeasily render senseless by other people on the road by driving recklessly or causing collisions.

In other words, the full potential of vehicle automation will only be realized when every car on the road is equipped with this technology. It will likely be several years before Level 5 automation is available to the public and functioning on most roads. Once available, it will almost certainly start out as a luxury option for the most expensive cars and take many more years to become widely offered, let alone mandatory. But we won’t see the roads become safer until every vehicle is able to drive itself and function completely predictably to any other automated vehicle. Once this happens, we will see the number of accidents and deaths on the roads decrease, maybe even to zero, and we will finally be able to safely look at our phones every day.