Director Of Google's Self-Driving Car Project Describes Extreme Lengths Taken To Avoid Another Bus Accident

Almost immediately after Google's self-driving car crashed into a public bus exactly one month ago, the company vowed that it would be taking steps to ensure that such an accident never occurs again.

Part of that process included Google releasing a statement about how it made necessary tweaks to its software.

"We've now reviewed this incident (and thousands of variations on it) in our simulator in detail and made refinements to our software," part of Google's statement said at the time. "From now on, our cars will more deeply understand that buses (and other large vehicles) are less likely to yield to us than other types of vehicles, and we hope to handle situations like this more gracefully in the future."

That being said, within that statement, we didn't quite know the lengths that Google went to ... until now. During Austin's SXSW festival, Chris Urmson, director of the company's Self-Driving Car Project, recently explained some of those steps that Google took to shore up the artificial intelligence of its autonomous cars ... and they're extreme to say the least.

Urmson said Google's team "implemented 3,500 new tests to make sure this won't happen again," as reported by CNET.

That's in addition to the 1.4 million miles of real-world testing that Google's fleet of autonomous cars have experienced thus far. Google's self-driving cars drive 10,000 miles a week, with the company also running three million miles of simulation driving in a computer program as well.

The hope is that through all this testing Google's deep-learning autonomous technology eventually sees it all, preparing the vehicles for literally anything they may encounter on the road.

As an attempt to show how Google's self-driving vehicle might be safer than human-driven cars, Urmson shared the story of a car traveling down a one-way street, approaching a four-way intersection, when its sensors detected a bicyclist and pedestrian amidst the traffic ahead. A bicyclist suddenly emerged, turning left the wrong way down the one-way street, but Google's autonomous car stopped and didn't hit the bicyclist. Impressive.

As we heard before about the bus accident, Urmson reiterated that Google's self-driving car saw the bus, making the incorrect assumption that the bus driver would slow down, with the bus driver making the incorrect assumption that the vehicle would stay put — neither happened, though, resulting in the accident.

ⓒ 2024 TECHTIMES.com All rights reserved. Do not reproduce without permission.
Join the Discussion
Real Time Analytics