Or, it could be radar. Or more lasers. We're not sure, except to say that it's new(ish), and that vision is presumably becoming more important for Google as they ask their cars to deal with more complex situations with more variables.
Another interesting tidbit in the update (posted by Chris Urmson, director of Google's Self-Driving Car Project) is the phrase about "teaching the car to drive more streets in Mountain View before we tackle another town." The Google cars can deal with changing environments and some level of dynamic uncertainty, but it needs a reliable basemap to use as a point of reference for lane width, traffic light formation, crosswalks, lane curvature, and more.
So, it might not be possible to tell them to drive you somewhere that they never been. To continue to speculate (because it's fun!), this might suggest how Google is planning on eventually making money on all of this: rather than making and selling autonomous cars, they'll maintain a continually updated database of road data that either car manufacturers or end users will have to subscribe to in order for their cars to operate autonomously.
Google's cars are still not ready for end users like you and I; they may be able to deal with 90 or 95 percent of situations autonomously, but that last 5 to 10 percent to make it to 100 percent autonomy (which is what's required) is probably as hard as all of the research, programming, experience, and machine learning that's gone into the cars up to this point.
It's impossible to plan ahead for every single scenario that an autonomous car might have to handle, so the key to unleashing the autonomous car is going to be a system that can learn and make decisions on the fly. In fact, it's almost certain that there will be accidents, because autonomous cars are still going to have to deal with all the other human drivers on the road (and also because no software is perfect).
But if we can get past our reluctance to place more trust in an autonomous system than we do in ourselves, autonomous cars have the potential to completely revolutionize our transportation infrastructure.
Get more from IEEE Spectrum
This article originally appeared on IEEE Spectrum; all rights reserved.