Nexar is partnering with Nevada public road transit authorities to create digital twins to reduce traffic and improve safety. In effect, crowdsourced dashcam image data is used to feed digital twins that represent virtual models of road work.
The partnership is another sign of consumer dashcam maker Nexar’s pivot to AI-infused digital twin-as-a-service offerings like its CityStream platform for governments and businesses.
“Leveraging vision, and in particular crowdsourcing from moving cameras roaming the cities, allows for a rich, live, and equitable digital twin that covers entire cities and not just high-traffic areas,” Nexar CEO Eran Shir told VentureBeat.
He said Nexar has developed AI algorithms to automatically extract road features from camera footage while still masking sensitive data. Such data is expected to feed digital twins that model city activity for civil engineering management.
The partnership with Southern Nevada’s Regional Transit Commission (RTC) partnership will weave real-time camera data into a comprehensive digital twin of the Las Vegas area that reflects the impact of work zones, changes in traffic signs, and road quality on traffic patterns.
A bet on digital twins
Government agencies in the city of Las Vegas and the state of Nevada are pursuing several such partnerships to apply IoT and SmartCity technology.
“This technology can help shape behavior and shift to a more proactive mindset by reducing the time in which problems are found, diagnosed, and fixed,” RTC engineering director John Peñuelas told VentureBeat.
Peñuelas observed that the RTC is the only agency in the country that oversees public transportation, traffic management, roadway funding, transportation planning, and regional planning efforts under one roof. The RTC works in collaboration with local governments across five cities and the Nevada Department of Transportation to plan and fund roadway projects and traffic signal operations.
RTC has been pioneering workflows for ingesting data from many different sources into a comprehensive digital twin of the city built on ESRI’s GIS platform. Now data from Nexar’s new CityStream service has been combined with existing fixed camera data, traffic sensors, and tracking RFIDs built into work zone traffic control elements.
The RTC authorizes hundreds of work zones each day and wants to track how they affect traffic. While these projects may all be under a permit, it’s not always possible to foresee the impact on traffic, safety, and other issues. RTC has seen how work zones affect bus lanes and bus stops and can track whether the zones are laid out safely. The new partnership can automatically show when work zone activity gets out of hand and causes congestion and other issues.
For its part, Nexar is also looking to tap into the autonomous vehicle market. Dashcams were purpose-built to capture data about accidents, and this vision data can help train AV systems on the corner cases of collisions, near-collisions, and other anomalous driving situations to learn proper road reactions.
Nexar is also pioneering AI collision reconstruction techniques to create forensically preserved digital twins of an accident scene. However, many in the industry still consider this type of application to be science fiction.
“Insurance professionals are still working on figuring out how to adapt processes to the rapid rise in camera use,” Shir said.
Crowdsourcing traffic in tomorrow’s metaverse
Nexar was founded in 2015 and has raised at least $100 million in funding to date. The company uses consumer-grade dash cameras to generate a fresh, high-quality, street-level view of the world and transient changes based on crowdsourced vision data from car cameras. Users capture about 130 million miles of road data per month. The company has pivoted beyond its initial service to provide digital twins for road planning and repair, delivery optimization, insurance, and autonomous vehicle training.
Video is not as fine-grained as lidar data, but it is easier to collect frequently. In some corners, video data is getting a second look as lidars encounter other issues.
In the long run, creating a metaverse that improves traffic will require improved collaboration and data sharing across various parties, including consumers, cities, car companies, and mapping services. Today this is something of a barrier to progress. Nexar, Cisco, and University of Catalonia researchers have proposed one IETF standard to help cars share digital twin representations of road conditions, traffic, and falling debris.