It isn’t just Ford’s (News - Alert) car plan that implies that we’re stupid, it is also the statistics surrounding automobile accidents. For instance, according to CrashStats, about 18 percent of injury crashes, and 16 percent of all police reported crashes were caused largely because we weren’t actually watching the road. And, you know that most folks don’t actually admit they were mucking with their cell phone when they ended up in an accident. Around 33K people die every year in vehicle related accidents in the U.S.; this compares unfavorably to gun deaths, which seem to get all of the focus, running a third of that number. Driverless cars, those that leave the driver out of the loop, are estimated to be able to drop the number of traffic accidents by 90 percent, which would save nearly 30K lives per year.
What we don’t talk a lot about is how much damage a human driver could do in an autonomous car driving world. And it is a lot because, much like you wouldn’t do well trying to compete with a computer in manufacturing, computers are vastly better drivers than we are; the problem is that they can’t predict what you’ll do. This all goes to why Ford, unlike most other car companies which are utilizing a concept Toyota called Guardian Angel, is going straight to having cars with no user controls. Because they’ll be a ton safer. And they are targeting Uber, which is kind of strange given Uber is using Volvos for their trial.
Let’s revisit autonomous cars again this week.
Chauffer vs. Guardian Angel Modes
There are several methods for ranking self-driving cars, but the one that stuck with me was provided by Toyota at the NVIDIA GPU Developer conference earlier this year. Basically they pointed out there were two methods under development, and it is likely some car makers, like Toyota, will have both.
Guardian Angel is likely the most attractive for drivers. This system is always on and in a car with drive by wire, where the driver input goes through the computer and can’t directly change what the car is doing. This is kind of like the cars you can “drive” at amusement parks. Disneyland is a good example where the cars are on a track which allows the feeling of self-control but no ability to drive off the road or do anything unsafe. If the car senses you are doing something stupid it simply, and seamlessly, takes over and you are suddenly a passenger. While this often reminds me of this scene from iRobot where Will Smith takes over from the computer, in reality it would work the exact opposite. Though, granted, if the system decided it wanted to kill you, you’d have little recourse (which is why a lot of us are a tad concerned about security in cars right now).
Chauffer is the simpler, less expensive and likely safer technology. This quite simply is where the car has no controls; it is effectively an elevator with wheels. You get into the car, tell it where you want to go and it takes you there. This reminds me of a very different movie clip from Total Recall regarding Johnny Cab; this is actually pretty close to where we are going, with the exception of the animatronic head (though that could come later to make the thing friendlier; hopefully in that case the result would be a tad less creepy). This is the path that Ford is on, and almost exactly so, because the target customer is a service like Uber.
One of the interesting back stories to this is that insurance companies are in a bit of a panic because a 90 percent reduction in accidents should result in a 90 percent reduction in premiums. That is a lot out of your P&L. But the good news is that human drivers on the road with self-driving cars will likely see their premiums go up massively because, given the close tolerances between robotic cars, one human has the potential to take out an impressive number of robotic vehicles by making a mistake. This suggests that human drivers may not last that long once this technology is viable.
Ford says they’ll have their solution on the road in 2021. Meanwhile, Uber has test vehicles—without the full driving capability— on the road this year. Once the cars are in use, it should take between five and 10 years to eliminate most vehicles currently being driven by people and replace them with robotic cars. Trucks, buses and Uber-like cabs will likely go first. This means that in about 15 years virtually every car on the road today will not only be obsolete but likely barred from driving on high speed public roads. The end result will be around 30K less deaths in the U.S. Given Virtual Reality should also be common well before then, that means most of us will be getting our spirited driving on a VR system perhaps at the same time a real car is driving us to our destination far more safely. Then again, if they don’t get their arms around that ugly security problem, we may all be looking at some variant of Will Smith’s experience; so let’s hope they get security sorted.
But, in the end, since we’ve shown that most of us can’t drive safely, it was only a matter of time before someone figured out that the best safety solution is getting humans out of the driving seat. Kind of ironic, given Ford took the car mainstream, that it will be Ford that is stepping out to kill the car as we know it and that this move will keep more of us alive. Go figure?