Will The Next Generation Even Need a Driver’s License?

It seems crazy to even be asking this question because driving has been such an integral part of at least American culture since the automobile began to be mass produced in the early twentieth century.  Driving has represented freedom and independence to teenagers for nearly a hundred years, however, we are now at the point where this question naturally arises. Of course, I’m talking about the possibility of commercially available self-driving vehicles.  As with most technologies, the idea of a self-driving car is not new (check out this article for a detailed history), but with the advent of GPS, lidar, machine learning (keep your eyes open for a post about machine learning later on), as well as other technologies, the self-driving future has entered into the realm of possibility.  In fact, self-driving cars already work, they just need to win over the trust of lawmakers and consumers alike in order to be part of everyday experience.

Let’s take a look at how a self-driving car is able to work.  The most important technology for a self-driving car is Lidar.  Lidar is essentially laser radar; it collects information about the distances to objects that allows the car’s computer to construct an accurate 3-D model of the world around it.  Tesla’s autopilot mode does this with sonar (like a bat), leveraging an ultrasonic array to achieve similar information about the car’s surroundings. Lidar is better for more accurate 3-D models, but sonar is significantly easier, cheaper, and doesn’t force you to put an ugly, spinning hat on the top of your car (if you want to learn more about lidar, here is an awesome article).  A self driving car needs much more than lidar to operate, though. Another imperative feature is a camera array, which gives the cars 360° color vision, important for obvious reasons.  Cameras allow the car to see the actual road surrounding it, so it can see how close it is to the lines or the curb. Cameras also give context to the model created by the lidar, sonar, or radar.  They allow the car to see the difference between a stop sign and a yield sign, whether a light is green or red, or what the speed limit is (if it didn’t get that from GPS already). Which brings me to the last super important hardware implementation, GPS.  Obviously, a driverless car needs GPS to know where it is, where it’s going, and where to turn. GPS is what would actually give the car the ability to navigate without a driver. Then the car needs to employ complex algorithms in a sophisticated computer system to analyze all the data that is coming in and react accordingly.  

Many companies have tried to enter the self driving car market, but perhaps the first was Google.  Their project Waymo (independent as of 2016) opened in 2009, and has been pioneering self-driving research and development for the decade since.  They achieved the first fully self-driven ride (no steering wheel, pedals, or technician) on public roads back in 2015. They also introduced the first car built fully around self-driving, nicknamed Firefly, in 2015 as well.  Before that (and still today because it’s easier) self-driving cars were just heavily modified production vehicles. Waymo started using the Toyota Prius, they’ve since partnered with the likes of Chrysler and Jaguar. Uber is partnered with Ford in a similar way.  Another important player in the driverless market is Tesla because although they don’t have anything near the like of Waymo or Uber, they rolled out their Autopilot mode a couple years ago, and now they claim that all the cars they produce have “full self-driving hardware,”  which essentially means they can get all the sensors in the car, but are years behind in the software development necessary to take advantage of that hardware.  To be fair, though, they did roll out autopilot, which boils down to being a really smart cruise control, to their vehicles in a software update and it is available to be used right now by any supported Tesla vehicle.  

Why are self-driving cars so important, anyway?  It’s more than just being a super cool thing for the public to marvel at (and it totally is).  Self-driving cars are the holy grail for automobile safety. The National Safety Council estimates that more than 40,000 people died in the US alone in automobile accidents in 2017.  In theory, self-driving cars could have prevented all of those deaths (check out this NPR article about human error in car accidents).  Huge problems of today like drunk driving, texting and driving, and distracted driving would all be eliminated by driverless vehicles.  Traffic slowdowns and the impact of transportation on the environment would both be severely reduced, as well. Forbes cites the Energy Information Administration as stating that self-driving cars could reduce fuel consumption for passenger vehicles by 44%.  

However, driverless technology is not perfect yet, and I very much doubt it ever will be.  For example, Uber pulled all of its self driving cars off the roads after one fatally hit a pedestrian on March 18th of this year.  Being from near Pittsburgh, I used to see their cars cruising through even my suburb regularly. It’s even worse that this happened because Uber required a technician to be in the driver’s seat ready to take over at any minute, but the one involved in the collision had become distracted.  The New York Times reports that Uber is unsure of how to proceed with self-driving in the future.  Despite crashes like Uber’s or the handful that have occurred under the watch of Tesla’s autopilot, self-driving cars must be the future.  Yes, there will be crashes that get huge press and make driverless technology look immature and scary, but we owe it to those 40,000 people from 2017 alone to make the self-driving future a reality.

Check out this awesome video about why self-driving cars need to hit the road ASAP.