Tech Tuesday: Autonomous Vehicles
Anyone who dares to take a quick survey of the press surrounding autonomous vehicles, alternatively called driver-less cars, will come away with a pleasant, gooey, hopeful sensation dripping down the back of their brain stem. That’s because most of the stories churned out by the major stakeholders revolve around some kind of utopian vision, one where all your problems are easily solved, cars cost nothing, the planet and everyone on it is saved, and your dog is taught Portuguese.
The flip side of this is the rally call of the luddites, which goes something like- “Look out! The machines are coming, and they’re gunning for your driver’s license! Probably your children, too!”
But as always, you can find the reality caught somewhere in the middle. The potential benefits of autonomous vehicles are many, and do include some truly transformative improvements to the way people get around. But there has yet to be a technology created that didn’t have some kind of negative repercussion.
At this very moment, there are production cars all around the world that sport some kind of autonomous feature, making them “partially” autonomous. Recently, this has included stuff like adaptive cruise control, which automatically adjusts your vehicle’s speed based on the distance to the car in front of you, or blind-spot monitoring, which signals the presence of a car the driver might not see when changing lanes. There’s also emergency braking, which applies maximum stopping power when an imminent collision is detected, saving precious fractions of a second over human reaction time.
But there are autonomous features even older than these. Electronic traction control, and even automatic transmissions could be considered forms of autonomous driving.
Of course, the Holy Grail for this tech would be a “fully” autonomous vehicle that requires zero human driver input, and there are currently a profusion of major automakers and technology companies racing toward that very goal.
Audi seems to be the most vocal when it comes to announcing progress updates. Most recently, the German marque sent a self-piloted A7 from Silicon Valley, California, to the Consumer Electronics Show in Las Vegas, Nevada, traversing a distance of 550 miles on public roads mostly driver-free. Other big names include Mercedes, Ford, and BMW, with tech giants Google, LG, and NVIDIA all getting in on the action as well.
Clearly, fully autonomous vehicles are just now peeking over the horizon. But the real question is: why should you care?
Click past the jump to learn about autonomous vehicles.
How It Works
Fully autonomous vehicles must harmoniously integrate a wide range of technologies to successfully navigate from point A to point B. While each tech is not individually groundbreaking, it’s the way in which they combine that represent the huge step forward.
First, an array of sensors is externally mounted to enable the vehicle to “see” what’s immediately around it. This includes radar similar to that found in adaptive cruise control, Lidar laser pulses to light up reflective surfaces, ultrasonic detectors for close proximity objects, and cameras to get an overall view of the road. These systems are used to pick out lane designators, other vehicles, road signs, traffic lights, pedestrians, hand signals, and just about anything else of importance.
Next, the vehicle must “know” where it is in order to correctly get where it needs to be. Therefore, GPS, gyroscopes, accelerometers, and altimeters accurately deduce location for large-scale navigation purposes, i.e., which freeway to take for the shortest overall travel time to Grandma’s house.
This torrent of rapidly fluctuating data is then fed to highly complex computer algorithms that string it all together and properly manipulate inputs for steering, throttle, and brakes. The system must also be adaptive, “learning” about the vehicle’s surroundings as it goes, updating itself to varying traffic patterns, road conditions, out-of-date maps, and other similar functions.
Finding the possible benefits of an up-and-coming technology is very easy. Just go to those who stand to make money from its implementation and ask them about it. I’m not implying the things they say are false, but rather would like to emphasize that until they are verified, they remain mere possibility.
With that being said, fully autonomous vehicles could very well do a whole lot of good. The greatest (and certainly the most commonly referenced) value is drastically improved long-term road safety. According to the World Health Organization, an estimated 1.24 million people die in traffic accidents every year, with the vast majority due to some form of human error. A computer algorithm won’t speed, run red lights, drive intoxicated, get distracted, drive tired, or otherwise muck it up behind the wheel. This would have the added benefit of reducing insurance costs and the need for traffic police.
Presumably, there would also be a greater separation between those who want to drive fast and those who want to just get there. We’ve all seen that a-hole weaving through traffic, thinking he’s a race car driver. Clearly, the place for that sort of behavior is a track, not the highway-commute home from work, and this technology could help divide the two conflicting mentalities.
Fully autonomous vehicles could also provide an enormous amount of freedom to anyone unable to operate a car, such as the elderly and disabled. No longer would these individuals require assistance to complete essential errands, like picking up groceries or visiting the doctor. Grandma would have no reason to berate you for never visiting, because all she’d need to do is get into a seat and buckle up (although, she’ll probably berate you anyway).
An increase in efficiency is another potential advantage to handing the reigns over to a computer. With the incredible precision that only a machine could provide, fully autonomous vehicles could drastically increase fuel mileage thanks to perfectly linear throttle application. Additionally, instant reaction times would allow vehicles to travel much closer together at speed, reducing congestion, and cutting air resistance to improve mileage even further. Long-term costs may also drop, as ride share programs would be far easier to implement if a car could arrive at a destination without human input.
Finally, there’s the obvious increase in convenience and comfort. Fully autonomous vehicles would become objects of luxury, not machines to be piloted. Of course, that’s not necessarily always a benefit…
Often, the latent drawbacks of a technology are impossible to predict until it’s fully implemented and adopted. Nonetheless, there are some aspects of fully autonomous vehicles that are clear problem areas.
Worst of all are the interactions between fully autonomous vehicles and the legion of terrible, unpredictable human drivers currently on the road. In fact, a recent white paper from the University of Michigan Transportation Research Institute outlines how safety may actually degrade when the computer cars make it mainstream.
This is, however, inevitable. There must be some kind of transitional period, and any automaker considering offering a fully autonomous vehicle will need to acknowledge that.
There are also ethical concerns. One common question raised is a scenario wherein a computer must decide between striking a pedestrian and swerving into oncoming traffic. Most people like to think that these kinds of split-second decisions are best left to a human being capable of moral judgment. What’s more, the question of liability is particularly thorny. Finding who’s at fault in an incident involving humans and an algorithm will soon make the jump from sci-fi movies to your auto insurance firms.
While issues within the nebulous realm of ethics and morality remain uncertain, one thing that is without question is the effect fully automated vehicles will have on the delivery service industry. As with any new form of automation, jobs once trusted to the hands of a human will get axed in favor of the tireless machines. Truckers, cabbies, and anyone else who makes his or her living behind a wheel will suddenly feel the pressure of far looser job security.
There are plenty of short-term costs as well. The complex systems described above don’t come cheap, and when you add in the need for additional government regulation, it’s hard to say exactly when the cost savings will overtake the initial outlay.
And while the proponents of fully autonomous vehicles want you to think these systems are fail proof, anyone who’s ever accidently downloaded the wrong file can tell you they most certainly are not. Here’s one nightmare scenario to keep you awake at night: some terrorist organization manages to crack the autonomous vehicle grid, causing wanton destruction and chaos. Cars, trucks, buses, and everything else on wheels is instantly rendered either useless or lethal, and the nation screeches to a proverbial halt. Mass hysteria ensues. Apparently, it’s a scenario that the Department of Defense is already preparing for.
Sounds like the plot for a great novel, no? Regardless, the idea of passing on driving responsibilities to an AI is clearly a double-edged sword. Philosophically, cars will become even more of a “thing” that takes you places. The gulf between man and machine will widen, and for some people, that’s kind of sad. Once an expression of individuality, cars will become just another tool to get around, as anonymous and unremarkable as a bus pass.
It’s similar to the loss of the manual transmission, but on an even greater scale. Of course, those who still care enough to turn their own wheel and press their own pedals will still be able to hit a track or deserted stretch of windy road. I won’t pretend for a minute that autonomous vehicles will ever threaten that — the passion behind the people that enjoy such activities is just too great. Such strange individuals may get relegated to the fringes, but in truth, I doubt they’ll ever be destroyed.
Fully autonomous vehicles are coming, it’s just a question of when. The possible benefits that they represent are just too good to pass up, even with all the looming disasters. At this point, any attempt to label them as either “good” or “bad” is completely irrelevant, because they’re gonna happen regardless.
We will almost certainly see first implementation in highly congested urban areas. This is where they are needed most, and such a setting represents the greatest challenges to the technology. From the cities, the tech will expand outward.
We are currently experiencing a rapid transitional period, one where partially autonomous is urgently pushed towards fully autonomous. It’s a time when the future of the automobile will be decided. It’s certainly exciting, to say the least.
Should the transition go smoothly, we might see fully autonomous vehicles sweep the industry, hitting the road with complete public acceptance in a decade. Should things go less than planned, I expect that adoption period to double, or even triple.
So far, things are looking good for the fully autonomous vehicle. Google’s efforts boast an untarnished safety record, even with 700,000 miles clocked for its fleet of 10 autonomous vehicles. Whether or not future progress will be as uneventful remains to be seen.