Communications of the ACM
rban Jungle a Tough Challenge For Google's Autonomous Cars
Credit: The Poole Party
Google's Chris Urmson, who heads the company's self-driving car program, said at a recent vehicle automation conference the project is entering its most difficult phase so far: building vehicles with the intelligence to navigate the chaos of city streets. Urmson said city driving requires a great deal more of an autonomous vehicle than the ability to recognize large objects and navigate around them. "You need to be able to deal with things like temporary construction," he said.
Urmson said the Google research team is focusing on "understanding the semantic meaning of the world." Google wants its vehicles to be able to recognize road signs and understand any number of social queues that shape the way people interact on the road, from the make of a given car to who is driving it.
Despite Google's ambitions and the great deal of excitement surrounding vehicle automation, conference attendees were conservative in their predictions of when such technology would find its way to market. More than half said it would be 2030 at the earliest before they trusted an automated car to take their child to school, with one in 10 saying they never would. Many said for the foreseeable future, self-driving vehicles would likely be limited to very controlled settings such as campuses or construction sites.
From Technology Review
View Full Article
Abstracts Copyright © 2014 Information Inc., Bethesda, Maryland, USA
No entries found