Last month Civil Maps, a developer of cognition systems for autonomous cars, launched a new low-cost 3D mapping system called the Atlas DevKit. The system consists of different sensors, including lidar, measurement systems, and communication devices, as well as the software needed to create 3D maps.
The Atlas DevKit’s hardware mounts on a car’s roof rack, where it enables a car to localize itself in six dimensions with the help of an ARM processor and generate a 3D map. Then, the software receives the sensor data and compresses it from gigabytes down to kilobytes, which enables the system to work as a crowdsourcing platform.
This means that the information autonomous cars need to navigate safely is generated faster and on a bigger scale. In fact, Atlas DevKit users can share their own maps directly to a cloud.
One of the biggest benefits is the system’s affordability. While many low-level localization and data collection systems range from $100,000 to $500,000, the Atlas DevKit has a starting point of $20,000 when you enter into an R&D contract with the company.
Furthermore, the Atlas DevKit is also sensor-agnostic, which allows customers to easily integrate the technology into existing systems and start mapping immediately. If you, or your car, already have what’s needed for the software to work, Civil Maps’ also offers an even more affordable, and software-only package: the Atlas Lite DevKit. Civil Maps was unable to present us with a price at this point, but confirmed that “it does come at a significantly reduced price.”
Civil Maps said, “Self-driving cars need highly accurate, 3D, constantly refreshed maps to understand their environment and safely navigate. Towards that goal, they need technology that can also scale.“
The Atlas DevKit is just the starting point. Civil Maps expects design and functionality changes as the project grows, and also hardware components, such as lidar, to become more affordable, making the technology even more accessible. The system is already available to qualified customers.