By Emma Bailey
In the Coronado National Forest, high above the saguaro cacti and desert yucca of southern Arizona, Forest Road 42 invites the intrepid motorist to cross creeks, dare winter snows and scale mountain peaks. Thousands of miles to the east, in Jacksonville, Florida, a state-of-the-art arterial highway system leads beach-goers and state senators through sprawling suburbs and across saltwater marshes.
Today, the same car, like a Ford F-150 or Buick Enclave, might see service in both areas. Tomorrow, at least if Google has its way, that may all change.
As of May 2015, Google's fleet of self-driving cars collectively logged more than 1 million miles. During the course of all that driving, they sustained only 12 minor accidents, though "not once was the self-driving car the cause," says program director Chris Urmson.
No one can ignore a car that's been driving itself for the human equivalent of 67 years. On June 25, Google announced the official arrival of its adorable, autonomous vehicles on the public roads of Mountain View, California. This summer the company plans to test its in-house built prototype, a two-seat, koala-bear-cute car equipped with a removable steering wheel, removable accelerator, removable brake and removable human.
Unlike its peers, who will incessantly gripe at its 25 mph maximum speed limit, the Google test car will not drive intoxicated. It will not respond to text messages. It will not fall asleep. It will not daydream about Chris Hemsworth. It will neither respond to raised fingers nor crush roving box turtles.
For that is the dream: a self-driving automobile with the reliability of a calculator. It has been estimated by the U.S. National Highway Traffic Safety Administration that 81 percent of car crashes are caused by human error. Could not those lives, that death lottery, be saved?
Other dreams fall in line. Some 45 percent of disabled American citizens still work. Autonomous cars could empower the other 55 percent. Costs of shipping would plummet as trucking companies lay off drivers. Valets and traffic jams would become as outdated as washboards and VCRs. If 20 percent of all cars were driverless, suggests one source, 724 million gallons of fuel would be saved, probably due to platooning and efficiency algorithms.
There are quagmires, to be sure. Casual automobile use may increase with access to technology, a variant of the traffic generation phenomenon. Expert Don MacKenzie, assistant professor of civil and environmental engineering at the University of Washington, said that automated driving pathways could lead to a 160 percent increase in automobile travel, bad news for the rapidly dwindling amount of available crude oil and natural gas.
And how would liability insurance be calculated? Who – or what – bears legal culpability in multi-vehicle collision? And all these dreams forget one thing: Forest Road 42 in the Coronado National Forest.
But first, the science:
A multitude of technologies work together to guide the Google driverless car. A GPS system informs the car of general routes and road regulations. A $70,000 LiDAR system, which uses a Velodyne 64-beam laser rotating 360 degrees and capturing 1.3 million readings per second, measures distances to surrounding objects. Stereo video cameras create overlapping fields of vision that record and follow road lines. Radar measures the speed of other cars. Altogether, approximately 1 gigabyte of data is gathered per second and fed into the car's brain to render a three-dimensional map.
Yet the system falters in heavy rains or snow. It flails on unimproved surfaces where road markings – sometimes even a road – do not exist. It cannot identify human hand signals.
Granted, most of the current obstacles to autonomous vehicle adoption can be surmounted. Vehicle-to-vehicle WiFi communication can prevent collisions. RFID chips in road signs can alert driverless cars to icy bridges and construction zones. Of course, these solutions present their own problems. You can hack a WiFi network, for instance. Unless you're a mega-powerful super-alien, you can't hack the human brain.
Where the road of autonomous automobiles will go, no one yet knows. But there will always be places, lost dirt roads and forgotten gravel pathways, where driverless cars will be unknown – because a human, unlike an intelligent automobile, can imagine what's around the next bend.
Image credit: Flickr/Roman Boed
Emma Bailey is a freelance writer and blogger from the Midwest. After going to college in Florida she relocated to Chicago, where she now lives with a roommate and two rabbits. She primarily covers entertainment topics and issues pertaining to the environment. Find her on Twitter @Emma_Bailey90