DeepMap Charts the Future of Automated Driving
Perhaps you think the world is already sufficiently mapped. With the advent of satellite images and Google Street View, it seems like every square inch of the globe is represented in data. But for autonomous vehicles, much of the world is uncharted territory. That’s because the maps designed for humans “can’t be consumed by robots,” says Tom Wang, the director of engineering at DeepMap, a Palo Alto startup that provides HD maps for self-driving cars.
These new kinds of vehicles need “a different technology of maps, a higher precision,” Wang says, because “the traditional map can be off by meters.” That is clearly unsafe for a self-driving car, which doesn’t have human intuition to tell it when to stop, for example. What’s more, for an AI to safely pilot a car on the road, “you need 3D information, you need a lot richer semantics, things like the traffic signals, a lot of different traffic signs, driving boundaries, connecting lanes,” he explains. “That information comes naturally for humans, but it is missing in the traditional map.”
What DeepMap does is twofold. According to Wang, they provide both HD mapping services and localization services. They are “integrated together to provide a high precision, like down-to-a-centimeter precision map of localization.” This demands a high volume of data to maintain both precision accuracy and localization accuracy—road conditions change minute-to-minute, so the maps guiding self-driving cars will have to update in real time as well. The end goal of mapping the entire world at the level of resolution necessary for self-driving cars to replace humans may be a long ways away. But with DeepMap, the road there looks a lot clearer.