Laura Bliss is CityLab’s west coast bureau chief. She also authors MapLab, a biweekly newsletter about maps (subscribe here). Her work has appeared in the New York Times, The Atlantic, Los Angeles magazine, and beyond.
The race to build navigation systems for driverless vehicles just heated up.
It’s all but certain that autonomous vehicles are the future of driving. But to get there, we’re going to need a better map.
That map will need to do two things: On the technical side, it needs to be able to talk to the vehicle in computer-speak about where to go, in what gear, and how fast. On the design side, it should be as intuitive, fun, and useful to a human being as the best app on her smartphone.
Mapbox, the firm that creates map software using mainly open-source data, announced Wednesday its official bid to build that map, with a new product called Mapbox Drive. It’s a map with the ability to direct a semi-autonomous vehicle with enough accuracy to make tight lane changes, and offer real-time traffic and road condition updates to the user.
Mapbox Drive has a model unique among car navigation systems. It doesn’t require a satellite connection, for one. Instead, it’s powered by millions of miles of telemetry data—that is, latitude, longitude, time, and elevation data, which help detect key details like lane width, speed, and turn restrictions. This data is scraped from other users of Drive as well as other Mapbox products, which are at the foundation of many mobile apps. This data is anonymized and aggregated, then processed (via satellite, wi-fi or cell connection) in real-time into Mapbox Drive, and aligned with the movements of the individual car. This data guides the car almost like rails do for a train, and it gets visualized alongside traffic info on the dashboard for the human driver—who, in a semi-autonomous car, can take over the controls at any time.
Eventually, Mapbox’s goal is to develop a system powerful enough to direct fully autonomous vehicles. And that’s feasible partly because, much like Waze, the map will keep improving itself, says Mapbox CEO Eric Gundersen. “The more people that use our maps, the more data that comes back, and the better the map is,” he says. “And then more people use our maps.”
According to Gundersen, Mapbox has already landed one contract with a major, unnamed auto manufacturer, and Drive will be live in one of their cars by the end of this year. As with all of Mapbox’s other products, the idea is to sell the Drive platform to multiple car manufacturers, who can brand and tailor it according to their needs and desires. “We’re the platform for platforms,” says Gundersen. This is a key distinction between the Mapbox philosophy and that of many competitors, including Google and Apple.
Google and Apple already have their own powerful digital maps, of course. Apart from their vast resources, that’s a big part of what sets these tech giants ahead of traditional auto companies in developing autonomous cars. But most major manufacturers are developing their own prototypes, too, and some are already rolling semi-autonomous models onto the market. Those automakers need great maps, and many would prefer to use a Google competitor that gives them greater control over the product, plus lower costs. This scramble for the right driverless-car mapping tech has fueled the recent explosion of market interest in cartographic data—exemplified in last summer’s $3 billion sale of HERE maps by Nokia to a group of German car makers.
So it is no surprise that Mapbox has pulled up a seat at the table. With their excellent track record in the mobile app world, the company could have a pretty good shot of getting their maps into your next dashboard. At the same time, the auto industry moves at a much different pace than the tech world does, and the competition is fierce. Anyone invested in the future of transportation should keep their eye on this map—and the others we’re guaranteed to see not too far down the road.