As recently as 2009, a driverless car seemed like a sci-fi fantasy — heck, why not make it fly or swim while you’re at it? Then, in 2010, the folks at Google revealed that they had been working on a self-driving car, and suddenly, the future was not only here, but it might even be cruising along in the lane beside you.
300,000 testing miles later, Google now projects that they will be able to bring self-driving technology to the market by 2018. Several car manufacturers (including Toyota, Audi, and BMW) have also been developing their own driverless vehicles, and self-driving cars are already street-legal in California, Nevada, and Florida. But are we getting ahead of ourselves? We aren’t talking about a vehicle on a track or in the relatively roomy space of the sky. We’re talking about free navigation of city streets and highways with all their attendant hazards (accidents, road work, pedestrians, other drivers). I think of San Francisco, with its wacky rules about left turns, one-way streets that suddenly switch directions, and ubiquitous double-parkers, and I wish the poor driverless cars luck.
The technology of self-driving cars
As we discussed in an earlier post, Google’s self-driving car relies on a combination of sensors (including lasers, radar, GPS, and a car-mounted camera) to get around. The car synthesizes all this data in real time and uses artificial intelligence to make decisions while it’s in motion.
On top of that, cars in general are becoming increasingly capable of communicating with one another. So when they’re equipped with high-speed broadband (as some models will soon be), they could potentially receive alerts about road conditions from other wired-up cars.
The developers of these vehicles envision a more efficient world, where autonomous cars could be packed together in “road trains” and multiple passengers could share a single self-driving car (like Zipcar, only it comes to you). Passengers could then theoretically spend their commute reading, sleeping, surfing the web, talking, texting, and relaxing.
Self-driving cars: the bad news
Google’s car can do a lot of things — keep a safe driving distance, find a parking space, brake to avoid a collision — but as of mid-May, it still couldn’t obey road signs or handle poor weather conditions (and there are conflicting reports about whether it can recognize pedestrians). Clearly, there’s still a long way to go before the cars are fully autonomous and ready for prime time … or rush hour.
Additionally, a self-driving car’s ability to get where it needs to go depends on a highly detailed, error-free map of roads and signals. This mapping already exists in GPS programs like Google Maps, and it’s getting more sophisticated all the time, but occasionally it’s just plain wrong (like when Google Maps sent me down a pedestrian-only cobblestone street in Ljubljana, Slovenia. That was fun.).
The National Highway Traffic Safety Administration (NHTSA) seems to feel the technology is far from ready. In fact, they recently called for a ban on the use of automated cars for purposes other than testing until they have conducted a thorough study.
Self-driving cars: the good news
The logistics of a successful driverless car may seem hopelessly complicated, but many cars are already available with “assisted driving” features like autonomous braking, self-parking, and sensors that warn you if you’re getting too close to a car or obstacle. So far, results are promising. Cars with autonomous braking, for example, had a substantial decrease in collision claims. Obviously, there’s a big difference between automatic brakes and a totally driverless car, but that gap is closing every day.
Self-driving cars also have another advantage over human drivers: they don’t get tired, distracted, or angry. According to the National Institute of Health, more than 90 percent of car accidents are caused by human error. Once the current technology issues are resolved, driverless cars have the potential to drastically reduce road accidents. These cars also offer mobility for people with health issues (such as impaired vision) that prevent them from driving standard cars.
There’s safety in numbers too. In a way, the more driverless cars are on the road, the better driving conditions will get. Cars will be increasingly more connected to each other and to government traffic systems, reducing congestion, accidents, and thefts. Also, the vehicles will learn as they drive and, using cloud technology, report everything from route information to road conditions, which can be accessed by other cars with the same software.
So how safe are self-driving cars?
Well, right now, it’s too soon to say. But if the technology continues to improve, we might be far safer in driverless cars, on roads filled with other autonomous vehicles, than we ever were with human drivers.
What’s your opinion? Would you trust a self-driving car to chauffeur you around?