X
Innovation

Google reveals how self-driving car works [video]

Google reps recently gave a filmed presentation that explained the technology behind autonomous vehicles.
Written by Tuan Nguyen, Contributor

When word spread that Google was testing a self-driving car, the technology was heralded as the  transportation wave of the future. Programmed with optimal fuel efficiency and safety in mind, the company claimed that it can reduce car accident by half. But despite some nifty navigational chops, beneath it all, isn't it still a heartless, calculating robot making some potentially life or death decisions?

It's these kinds of concerns that was a cause for alarm for some when it was reported that Google was lobbying the state of Nevada to bestow driver-less cars with street legal status. And it was only a couple months later that such fears were realized when one of their test cars collided with a street vehicle driving nearby the company's Mountain View, California headquarters. Google's response to the accident was to basically say that it wasn't their fault.

In an email, a Google spokesperson told the Mountain View Voice, "We regret that a Google driver recently caused a minor accident, and we're grateful that no one was hurt...Safety is our top priority. One of our goals is to prevent fender-benders like this one, which occurred while a person was manually driving the car."

And perhaps to assuage any heightened anxieties the public might have about cruising alongside fully autonomous automobiles, the company recently gave a filmed presentation at the IEEE International Conference on Intelligent Robots and Systems in San Francisco in which they go through an in-depth nuts and bolts explanation of how the technology works.

During the roughly 15 minute talk, project head Sebastian Thrun and Google engineer Chris Urmson detailed the inner workings of the car's roof-mounted range finder, a Velodyne 64-beam laser device in which the navigation system is built on.

Other features include:

  • Front and rear radar systems that helps the vehicle calibrate its movements based on the speed and positioning of cars on the freeway.
  • A camera located near the rear-view mirror that reads traffic lights.
  • A GPS inertial measurement unit and wheel encoder that keeps track of the vehicle's changing coordinates.

Google also gave the audience an insiders look at what allows them to feel confident about the safety of the technology: a "multi-layered" approach that combines GPS with exquisitely detailed maps and calculations based on comparisons between real-time and previously recorded data on road conditions. Basically, the computer is in a constant state of ensuring that the sensory information it's receiving is verifiably correct. With all that available data, the computer can take advantage of a command system that can react faster than people to hazards and changing road conditions.

Whether this is enough to make drivers any more comfortable about sharing the road with drones is still something that needs to play out. At the very least, though, assurances are going beyond a earlier stance that amounted to nothing more than blog posts essentially said "just trust us."

(via Discovery News)

Related:

More care-safety technologies:

This post was originally published on Smartplanet.com

Editorial standards