When the Google autonomous car made headlines around the globe a few months ago, everyone started thinking about a brave new automotive world. A world in which our cars drive themselves, eliminating traffic accidents and freeing us to practice yoga or think great thoughts during our morning commute. For proponents of autonomous cars, the future was a bright, shiny utopia.
Google co-founder Sergey Brin forecast in 2012 that self-driving cars would be ready in five years. But author Lee Gomes has looked in to what’s next for the Google car and he tells us, in an article for Slate magazine, “Don’t hold your breath.” That’s because the Google car depends on maps – incredibly detailed maps that contain the exact three-dimensional location of street lights, stop signs, crosswalks, lane markings, and every other crucial aspect of a roadway. To create such maps, a dedicated vehicle outfitted with a bank of sensors scans the roadway repeatedly. The data is then downloaded and every square foot of the landscape is pored over by both humans and computers to make sure that all important real world objects have been captured.
Before Google’s vision of a future filled with autonomous cars can be realized, all 4 million miles of U.S. public roads will have to be mapped in this intensive, time consuming and expensive way. So far, only a few thousand miles of road around Google headquarters in Mountain View, California have been charted. The company boasts that its autonomous cars have driven more than 700,000 miles in complete safety, but those are the same few thousand mapped miles, driven over and over again.
Problems occur when something changes in that carefully plotted environment. A stop sign gets taken down and a new traffic light installed in its place. Or construction causes traffic to thread its way through a temporary detour marked by orange cones. If those things are not on the map, the computer in the Google car can’t see them, with potentially dire consequences.
Where I live, a major interchange between two interstate highways has been altered dramatically. What was once an exit from the left lane to a bridge over a river has now become an exit from the right lane leading to a bridge over the same river but a half mile further south. If my car was being driven by a computer with an outdated map stored in its memory, it would have me crashing through a guardrail where the old exit used to be and landing upside down in the water. Which would be unsettling, to say the least.
To be successful, the Google car and all other autonomous cars will need to get beyond the detailed map control system and conquer the realm of artificial intelligence, popularly known as AI. In that world, the electronic control system will be able to tell the difference between the flashing lights on a tow truck from those on an ambulance. It will “see” pot holes and stop signs. It will know if a traffic light is red or green. And it will be able to predict the path and speed of pedestrians and other vehicles. It’s not that any of this is impossible; it is that the artificial intelligence is a long way from being ready for prime time. In the meantime, automakers like Tesla and Mercedes are using radar and forward-facing cameras to “read” the road, which will allow the first generation of semi-autonomous cars to hopefully hit the road soon. The automakers seem to have a better handle on this autonomous car thing if you ask me.
“None of this reasoning will be inside computers any time soon,” says Raj Rajkumar, director of autonomous driving research at Carnegie-Mellon University, former home of both the current and prior directors of Google’s car project. Rajkumar adds that the Detroit car makers he collaborates with on autonomous vehicles believe the prospect of a fully self-driving car arriving anytime soon is “pure science fiction.”
None of this means that new technology like adaptive cruise control and automatic braking systems will not become more common in our cars. It just means that when we go over the river and through the woods to grandmother’s house, it will still be with a real human driver at the wheel for the foreseeable future.