Curtiss Motorcycles unveils electric concept prototype: Zeus
Cedars-Sinai study finds prolonged exposure to coarse PM in LA Basin air led to genetic changes in rat brains

Civil Maps moves to Arm processors; auto-grade, power-efficient, mapping and localization solutions for autonomous driving

Civil Maps, a developer of edge-based HD mapping and localization solutions for autonomous vehicles, has ported its vehicular cognition stack to run on the Arm family of processors. Arm is the leading licensor of automotive-grade system-on-chips (SoCs) and system-on-modules (SoMs), with more than 1,000 technology partners.

The move by Civil Maps will enable self-driving car developers to leverage an Arm SoC with Civil Maps embedded software for enabling vehicles to navigate safely and localize on the road. Currently, no other competing solutions, which commonly rely on x86 and/or GPU-based architectures, offer developers a path to production-scale vehicle autonomy, the company claimed.

The Arm-powered HD maps and localization solution, which further enhances the already streamlined, lightweight performance of the Civil Maps platform, will enable greater levels of reliability and efficiency while overcoming chronic limitations of power consumption, CPU, storage, and network requirements, the company said. The timing of the transition coincides with the automotive industry’s adoption of Arm technology for electronic control units (ECU) that manage an expanding range of in-vehicle systems. Civil Maps’ migration means that OEMs will be able to leverage the company’s technology without changing their existing supply chain.

Committing to an Arm-based architecture will greatly assist Civil Maps in seamlessly commercializing their technology to leading automakers. This approach will allow the OEMs and technology companies involved with autonomous development to implement a scalable localization platform with lean compute requirements.

—Tim Dawkins, Autonomous Car Specialist at SBD Automotive

As part of the transition, Civil Maps has optimized core components of its algorithms and software for the Arm CPU architecture, known for its high-volume manufacturing process and auto-grade certifications. The efficiency of Civil Maps’ exclusive Fingerprint Base Map technology allows the company to leverage the Arm Neural Network accelerators in these low-power, low-cost SoCs.

Founded in 2015, Civil Maps has received funding from Ford Motor Company, SAIC, Motus Ventures, the Stanford-StartX Fund, and other investors.



Humans can drive without HD maps, and in most cases, without maps at all.
Humans have driving and traffic rules and object recognition.
We can drive without Lidar and Radar, just high resolution colour stereo imaging and almost perfect object and scene recognition (it not perfect, then fit for purpose anyhow).
We do not need to know where everything on the route is in advance, just the few turns we may need to make.
It looks like a solution looking for a problem to me.

An electric car would be far better marketing than this 10 mpg gas guzzler.

The comments to this entry are closed.