Google “Mapping” Real World With Incredible Precision For Self-Driving Cars

A fascinating article in the Atlantic appeared this weekend. It’s mostly about Google’s self-driving cars and how they operate technically. But it’s also about something much bigger: how Google is now effectively “crawling” the real world as it has crawled the web for years.

The analogy or insight isn’t mine it’s the Atlantic’s. But it’s spot-on.

With its multiple “real world” initiatives (Street View, Project Tango, Indoor Maps, Android location and others) Google is building a virtual representation of the real world that can be used for many and varied purposes, including offline ad tracking, a range of mobile consumer services and search improvements (Google Now).

The article asserts that Google is making the world “legible” to robots. Here’s an except that shows, however, Google’s self-driving cars are years away as a mainstream phenomenon:

The key to Google’s success has been that these cars aren’t forced to process an entire scene from scratch. Instead, their teams travel and map each road that the car will travel. And these are not any old maps. They are not even the rich, road-logic-filled maps of consumer-grade Google Maps.

They’re probably best thought of as ultra-precise digitizations of the physical world, all the way down to tiny details like the position and height of every single curb. A normal digital map would show a road intersection; these maps would have a precision measured in inches. 

But the “map” goes beyond what any of us know as a map. “Really, [our maps] are any geographic information that we can tell the car in advance to make its job easier,” explained Andrew Chatham, the Google self-driving car team’s mapping lead.

“We tell it how high the traffic signals are off the ground, the exact position of the curbs, so the car knows where not to drive,” he said. “We’d also include information that you can’t even see like implied speed limits.”

Google is thus generating and processing massive amounts of data — the height of traffic signals, the exact position of curbs — in order to make the world intelligible to self-driving cars (and beyond). To make self-driving cars work in locations other than Mountain View, California Google would accordingly need to obtain much more highly detailed data for essentially every street in the US and the world.

It’s “Street View 2.0” but at a level of almost unimaginable precision. And to work at scale, any time soon, Google needs to start outsourcing (to car makers and/or random citizens) the mapping function. I’m not clear that this is even possible but that’s what logic would dictate to achieve national and international coverage.

There are many scenarios in which self-driving cars would benefit individuals and maybe society as a whole (e.g., fewer accidents). However it’s very likely that Google will meet with strong resistance because the effort to map the world at this level of precision will fuel further Google paranoia.

Ultimately Google may need to “open source” the effort and bring multiple companies and even governments into the initiative in order to both accelerate the process and secure their approval or acquiescence.