Futurists and tech enthusiasts insist that self-driving cars are going to be a technological revolution, dramatically reducing the death toll and injury rates of vehicular collisions, but skeptics tend to believe that no automated process can control a car like a human driver. There’s still a lot we haven’t worked out with this up-and-coming technology, but the conceptual framework is already in place.
So how is it, exactly, that self-driving cars could reduce the rate of traffic accidents? And is it really feasible?
The Common Causes of Car Accidents
First, we need to understand the most common causes of car accidents, which include:
- Distracted driving, including texting and driving, which occurs when a driver takes their eyes off the road or when something else grabs their attention.
- Impaired driving, which happens when an ordinarily reasonable driver consumes alcohol or is affected by drugs that delay reaction time or distort perceptions.
- Adverse weather conditions, which are hard to prevent, and include snow and ice, fog, and heavy rain.
- Speeding, since the faster your vehicle travels, the less time you have to react to a sudden change in your environment.
- Driver error, such as turning the wrong way down a one-way street or blowing through a red light.
Assuming a self-driving car can “see” the world in a way similar to human beings, understand traffic laws the same way, and maintain consistent control of the vehicle, we can already throw out multiple causes of accidents. A self-driving car can never be impaired by alcohol or drugs, and can never be distracted like a human being can. Its error rates, again assuming a similar level of environmental understanding as an optimal human, would also be negligible, and self-driving vehicles would never travel above the speed limit. Adverse weather conditions may still be a problem, but after eliminating multiple categories of common accidents, a lower cumulative accident rate is practically guaranteed.
How Self-Driving Cars See the World
Now let’s examine whether a self-driving car can really see the world as fully as a human being, since this was an underlying assumption in our assessment. Self-driving cars are equipped with multiple methods of perceiving its environment, relying on motion sensors, RADAR, LiDAR, 360-degree camera views, GPS data, and internal 3D maps to form a clear picture of its surroundings. In actuality, self-driving cars may be better at perceiving their environments than human beings; not only can they see in a 360-degree sphere at all times, they also have multiple methods of detecting objects in their surroundings.
What We Know
The capabilities of self-driving cars aren’t just hypothetical, either; Waymo self-driving cars alone have driven more than 10 million miles on real roadways, which is more than 10 full lifetimes of human driving, yet there’s only been one high-profile collision—and it wasn’t the self-driving car’s fault. On top of that, their cars have driven more than 7 billion miles in simulations.
Real-world tests are showing pretty conclusively that our self-driving car technology is already at a level superior to human drivers, and engineers are continuing to make improvements.
That said, there are some real risks with self-driving cars. Because the software is fully automated, if a hacker were to gain control of its system, they could hypothetically control not only where it drives, but how—ultimately putting lives in harm’s way. Self-driving cars are equipped with multiple systems for sensing the environment, but if more than one of them is knocked out, they could become unable to reliably sense their environments. On top of that, self-driving cars’ autonomous and predictable procedures could leave them vulnerable to bullying by human beings.
Still, these security vulnerabilities aren’t that much different than what we face in normal cars. Instead of hacking into your car’s software, a criminal could cut your brake lines. Instead of knocking out a sensor, a human driver could be dealing with an unexpected weather condition, or a sudden medical event, like a stroke or heart attack. And human pedestrians can also bully human drivers in their cars—there’s just more of a threat of someone running you over.
The Bottom Line
There’s clear evidence, both in theory and in practice, that self-driving cars are already much better than human drivers, capable of reducing the rate of traffic accidents dramatically—and they’re only going to get better from here. We should also consider that self-driving cars don’t even have to be better—they could just be on par with the skill of human drivers—for us to accept them. And if they are capable of reducing the death and injury toll on our roadways, it’s not just a good idea to adopt them; it’s our moral imperative.
Waymo Jaguar I-Pace – by elisfkc on flickr – some rights reserved
Guest Author Bio
Jamie is a freelance writer who covers trends in business, technology, and health. She loves to go skiing, camping, and rock climbing with her family.