By Paul Drysch, VP of Sales and Business Development
The mapping world is gearing up for the autonomous revolution. Maps for humans are being replaced with maps for autonomous vehicles. Teaching a car to drive itself requires teaching it how to see and interpret the world around it using a new category of maps: HD maps for autonomous vehicles. The opportunity to help shape the future of automated driving is what drove me to join Civil Maps.
Civil Maps is developing a new mapping approach designed from the ground-up for autonomous driving. Unlike the maps you’ve used on your mobile devices, these maps are intended to be interpreted by machines. They are specialized maps that tell vehicles where they are and point out the things that are important. Civil Maps’ sensor fusion and voxel-based fingerprint localization technology provides vehicular cognition for self-driving cars. In simple terms, it’s like the difference between driving in your neighborhood versus driving for the first time in NYC during rush hour. In your neighborhood, your brain knows where everything is, where to go, and how best to get home. It’s easy. For companies using traditional methods of mapping and localization, every inch of a road is like driving in NYC for the first time.
With Civil Maps, it’s just the opposite. It’s like having someone in the car pointing out the signals and street signs as you drive, helping you every step of the way. While that may not seem essential for a human driver, teaching vehicles to see and interpret everything around them is critical for self-driving vehicles. There are several big advantages of Civil Maps’ approach:
- Affordable in-vehicle hardware: Interpreting every square inch of the world in real-time can be computationally expensive. Most vehicles attempting fully automated driving using this approach require tremendous processing power, energy consumption, and data storage that is just too cost-prohibitive for anything other than the small geofenced pilots you read about in the news. Civil Maps’ lightweight maps provide map content and the intelligence to tell cars where to prioritize attention while requiring considerably less processing power than traditional maps. An inexpensive consumer-grade CPU is all that’s needed to power an autonomous car when using Civil Maps, thus lowering the car’s build-of-materials (BOM) costs drastically.
- Most up-to-date vehicles: With Civil Maps’ proprietary fingerprinting technology, the map files themselves are tiny. It’s a bit counter-intuitive. For example, a competitor’s base map of San Francisco, which is about 4,000 kilometers, can take up to 4 terabytes of data. This is where Civil Maps Edge Mapping™ really shines: Our Fingerprint Base Map™ of the same area requires only 400 megabytes. Small maps are a game-changer. With small base maps, it’s possible for full map updates to be downloaded to cars over-the-air (OTA). To drive safely, autonomous vehicles need maps that are constantly updated. Vehicles using Civil Maps will have the most up-to-date maps via simple, secure, and inexpensive OTA updates.
- Map creation independence: Our mapping approach not only takes less computational power to interpret data from the sensors, it takes less effort to create the map. Autonomous vehicles using Civil Maps all become map creators, too. There’s no need to rely on expensive, specialized survey vehicles to create HD maps. Now every Civil Maps-enabled vehicle contributes to maintaining the map, e.g. true crowdsourcing. Civil Maps customers are able to contribute and receive credit for these map changes, reducing overall map costs required for autonomous-capable HD maps. And because the maps are so lightweight, everything can be done over the air.
Civil Maps offers a unified platform that provides for HD Semantic Map creation, map usage, centimeter-level localization, and sensor fusion integrating directly to the decision engine of the autonomous vehicle. Civil Maps already provides these services for nine different automotive OEMs, a number of major mapping companies, and is backed by OEMs like Ford and SAIC.
It’s an exciting time to be charting the world for autonomous vehicles. I’m excited to be a part of Civil Maps family because we have rethought the way maps are made and consumed by autonomous vehicles. We’re rewriting the map for autonomous driving!