One of the most interesting sessions at last month’s Embedded Linux Conference in San Francisco was a keynote presentation about the technologies, capabilities, and challenges associated with Google’s self-driving cars, which have now traversed some 400,000 miles on public roadways. Think of it as “Prius meets The Matrix.”

In the 25 minute video below, Andrew Chatham, a senior staff engineer at Google, examines Google’s self-driving cars from a system and software developer’s perspective. He describes various technical considerations and challenges, and shows what the car’s Ubuntu-powered onboard “brain” “sees” as it deftly navigates through its computer-generated 3D virtual environment — avoiding pedestrians who unexpectedly cross its path, dodging an oncoming truck on a narrow winding mountain road, and at one point getting trapped in an endless loop at a roundabout. What happens if a car mistakenly enters a freeway from an offramp?

Here’s the video of Chatham’s 25-minute Embedded Linux Conference keynote talk:

Below are a selection of slides from the talk, showing the self-driving car making its way through its computer-generated 3D virtual environment despite various obstacles.

(click thumbnails for larger images)

(click thumbnails for larger images)

In case you’re up for more, here’s Sebastian Thrun’s TED 2011 talk:

Additional ELC 2013 San Francisco talk videos are available here. Information on future Linux Foundation events may be found on the LF’s website.