A study released in February by the US DOT revealed a startling fact. A full 94% of crashes are the result of a human error. And, only 2% are caused by a vehicle malfunction or defect, which ties in perfectly with a landmark report from Malcolm Gladwell in the last New Yorker about safety recalls. It turns out we all need to pay attention to the road, set down the phone, and practice better driving habits.
Today, Google announced they are going to give us some help with that problem.
The “ladybug” version of their custom-built autonomous car, the one that replaces all of those Toyota Prius and Lexus RX450h SUV test cars, is going public. If you live in Northern California, you might even see one driving on public roads this summer.
In fact, Google intends to drive these cars in urban areas with a speed cap of 25 MPH. A human driver will still sit behind the driver’s seat, but we’ll have to come up with a new term for that spot in the vehicle — these cars have a removable steering wheel. Eventually, they might not need a gas pedal or brakes.
One of the goals if the new road test using a fleet of autonomous cars is to find out how congestion and construction impact drive routes and autonomous response. As I’ve covered several times before, the main hold-up in turning cars into robots it that we still don’t know how they will react to unusual situations. Can a “smart car” speed up and get out of the way of a minivan driving too fast on the highway? Can it swerve to avoid road kill?
Another impetus for this new road test is to gather information about public interest. (In some ways, that’s what Google’s been doing all along.) They want to find out how people would like to drive robot-enhanced cars. Do they use them for a boring commute? To drive up to the supermarket and get eggs late at night? When do they most need the augmentation? When will we let the car take full control?
Google claims they have extensively tested the new version of the autonomous car in private — seeing how it responds to bumps in the road, hot and cold climates, and other conditions. Private tests can be limiting, though. The company claims the next important step is seeing how the car performs around other cars and human drivers.
Mostly, this is an engineering exercise. There’s a boatload of precise math calculations that go into programming an autonomous car related to speed, position on the road, the angle of the car, and even variables like going on a controlled descent down a hill at high speeds versus slower speeds on a rainy day in July. That’s really why the testing takes such a long time compared to getting a car like a new Audi A3 (which debuted in the last year) ready for drivers to take the wheel.
Here’s hoping the new test leads to some amazing conclusions. If the goal is to reduce accidents even by 10% or 20%, I’m all for it.
This article is published as part of the IDG Contributor Network. Want to Join?