Self-driving cars seem to be in the news constantly these days. All sorts of questions are being raised about their suitability and whether or not they’re safer than human-operated vehicles and cars.
While these questions are a tad premature (the only autonomous vehicles on the road today are in testing), it’s understandable that people are concerned about the safety of such vehicles. But, is this concern valid?
We’re exploring where this industry is headed–and how it’s already impacting us–to get a better sense for what’s at stake.
What Self-Driving Cars, Agent Orange, and the Internet Have to Do With Each Other
When we think of self-driving cars today, we mostly think of Google, Uber, and Tesla, but did you know it was the government who jump-started the self-driving tech we utilize today? DARPA, the Defense Advanced Research Projects Agency is the arm of the United States Departments of Defence that develops new technologies.
Most of those technologies (like Agent Orange, the chemical defoliant used in Vietnam and now known to have deadly side effects) go on to be used exclusively by the military, but some–including the internet–are opened up for public collaboration and adoption.
Driverless cars are another DARPA initiative. In 2004, DARPA had exhausted millions of dollars on autonomous vehicle technology but was getting nowhere. So, it sponsored the Grand Challenge in the Mojave Desert in California.
The Race Begins
The competition was open to anyone–college teams, random groups of friends, companies, and more–but despite the multi-mile course the robotic cars were required to drive, the car that made it the furthest was one from Carnegie Mellon, which never even reached the finish line.
DARPA went on to host two more challenges in subsequent years, each more popular (and, luckily, more successful) than the last, but what the races really did was kickstart an industry. Collaboration and friendly competitive converged with newly available technologies to make a Jetsons-like future suddenly seem, well, possible.
The Competition Heats Up
Now, companies are jumping into the fray like mad. Waymo, Google’s self-driving spinoff, has long held the industry lead (the company wholesale hired one of the DARPA race’s winning teams), but Uber, Lyft, and others are hot on its tail.
All the big auto industries–Ford, General Motors, and more–are invested somehow, and even companies like Apple have indicated their involvement. Plus, a whole host of support companies have popped up, including laser, robotics, and computing companies.
Most people believe that in the same way the horseless carriage dramatically changed the American landscape, so self-driving cars are poised to usher in a second revolution in transportation and how we live.
How Can Self-Driving Cars Make the World Better?
While fear is a common reaction to self-driving cars (especially as more and more cars are tested, and there are more accidents), they’re not all bad. In fact, they stand to make our world quite a bit better. Skeptical? Let’s break it down!
One of the things that’s important to know about self-driving cars is that they are not single pieces of technology. In fact, they’re very complicated amalgamations of incredibly advanced systems. One way we can better understand them is by understanding some technology that’s already become a part of the cars we drive now.
Adaptive Cruise Control
Adaptive cruise control (ACC) is like cruise control, but better. Utilizing technologies similar to those found in self-driving cars, including cameras and lasers, ACC brakes or decelerates to keep a specified distance between you and the car in front of you.
It can fully replace driving as an autonomous car could, and most systems won’t brake entirely only slow down (and of course it can’t steer, either), but ACC is a taste of what’s to come in self-driving cars, and helps us understand how we might avoid more accidents in the future!
Automatic parking can refer to a variety of technologies, but typically it’s used to describe what Toyota first integrated into its Prius line in 2003, a car that can parallel park with zero or limited involvement from the driver.
This computer assistance mirrors the assistance that self-driving cars hope to provide one day by using sensors to gauge the distance between obstacles and parking accordingly, navigating safely around the curb, other cars, and so forth. Drivers are both freed up to focus on other hazards and assisted, so they avoid impact.
Lane Keep Assist
Lane keep assist (LKA) often works in conjunction with lane-departure warning (LDW) as another precursor technology to self-driving vehicles. While LDW warns you via seat vibrations or audible alarms, LKA helps you stay in the lane by steering you back in if you leave.
Again, lane keep assist isn’t self-driving, but it’s a great taste of what’s to come and another example of how technology can make us safer. It’s available today in Cadillacs, Hyundais, Teslas, Jeeps, Hondas, and more.
Traffic Sign Recognition
Finally, traffic sign recognition is yet another technology that gives us a taste of self-driving cars at work. Traffic sign recognition uses a variety of algorithms to identify traffic signs–especially speed limit signs–and provide the information to the driver.
This technology is common in high-end European vehicles.
As we consider the impact self-driving cars will have, and whether or not they’ll be safe, it’s important to remember the first car was a horseless carriage–it was defined by a lack of horse! The next revolution in transportation, then, will be defined by another lack, a lack of driver!
Problems With Self-Driving Cars
Despite all the appeal of self-driving cars, there are still some concerns. The software, for example, is far from perfect yet. One of the biggest challenges is that software has to identify and then predict what people and real-life objects will do. Humans, in particular, are incredibly difficult to predict.
While that unpredictability is one of the very things that makes self-driving cars appealing, it’s also one of the things that makes them so difficult to create. The self-driving Uber car that killed a woman in Arizona, for example, most likely had some of software glitch that failed to properly identify the pedestrian.
Beyond the Software
One of the biggest challenges for engineers working on self-driving cars is the software, but that’s not the only challenge. The car’s ability to “see” obstacles, signs, road conditions and more is a vital part of the whole unit.
Traditionally, cameras and radars have been used to mimic the vehicle’s eyes, but there are problems with both. Cameras only capture surrounding images and data in two dimensions and radars, while inexpensive, provide limited details.
Lidar, as it turns out, is the answer. Lidar is similar to radar in that it sends out pulses and reads the response to create a 3-D picture, but instead of radio waves, lidar uses pulses of light. The light pulses from a rapidly revolving sensor on the car’s roof.
A Different Radar
The downside, as we’re sure you can easily see, is that it’s not constant. The light doesn’t hit everything, and even when it does, the computer has to read and make instant deductions. For fast moving objects, such as a runner or a dog racing across the street, this can become a safety risk.
It’s also extremely expensive; usually twice the price of the car it’s attached to! The price isn’t a safety risk, but it’s clear that lidar has a ways to go before it’s where it needs to be for us to use and benefit from it in a self-driving car.
Even after millions of miles have been driven in self-driving cars, self-driving cars are still wrecking and still not ready for deployment by the general public. When they are rolled out, it probably won’t be as a car you can park in your garage.
Since it will be even longer before self-driving cars can handle diverse roadways and terrains, these won’t be cars you can take for weekend jaunts to the beach. Instead, they’ll likely be best suited at first for designated routes, like a daily commute that picks up and drops off at the same place every day.
Ultimately, self-driving cars have a long way to go. If they were to flood our streets tomorrow, there would be room for concern, but as it stands, they’re still undergoing intense scrutiny and testing.
We can feel confident that just like they’ve made huge strides in the last decade and a half, we’ll continue to see leaps forward over the next ten years. Here’s to safer roads with or without self-driving cars!