Together with KDDI and Oracle, Nexar Unveils AI Application that uses Crowd-Sourced Vision Data to Find Available Street Parking in Tokyo 

Supported by KDDI and Oracle Cloud Infrastructure, Nexar shows how Automotive Edge Services will work in the future and releases a video showing how a network of AI dash cams can spot empty spaces in real-time

Nexar, the leading computer vision AI company, released a video demonstrating the first ever AI-powered parking space-finder app. Demonstrated with Oracle & KDDIs' mobile and edge infrastructure, the app reviews dash cam footage to detect on-street free curb space in Tokyo's busy streets using edge compute environments. Nexar, KDDI and Oracle are part of the Automotive Edge Computing Consortium (AECC), which works to create the computing and connectivity infrastructure for connected cars of the future.

Nexar's app can dynamically and safely scale the sharing of transient conditions in near real-time. This demonstration comes on the heels of a similar real-time on-street parking app pilot in New York City, where thousands of Nexar's popular dash cams were driven around the city, capturing 220 million miles of driving each month and spotting empty parking spaces in real-time. Locations and photos are shared instantly to Nexar's app, so that a circling driver can quickly grab the empty space.

Using the power of AI and the distributed edge architecture, Nexar solves the issue of latency, and compute and connectivity costs by consolidating and verifying multiple vehicles at the network edge. Nexar's goal is to solve the geolocation class of problems: HDMap fresh change collection, intelligent driving hazards, provide blockages and convenience alerts, and create traffic scheduling cruise assist, using edge compute environment and resources. This way, Nexar can spot and map empty parking spaces "seen" by these dash cams as they drive through the city, sharing them through push notifications to other drivers to the free parking spot.

Tokyo's AECC pilot is a breakthrough in proving the viability of such applications for the future of 5G networks, using in-car camera SDKs to compute the Uber H3 spatial index, which is used to distribute data across multiple edge clouds including Oracle Cloud Infrastructure (OCI). The pilot also uses AECC aggregation and consolidation on the network edge. This PoC proves the feasibility of cost-effective scale and elasticity of automotive edge services that will drive different use cases leveraging the powerful 5G and edge compute industry trends. In addition, Toyota, DENSO, NTT, Ericsson and Intel are charter AECC members.

The parking-finder app used the AECC's distributed edge architecture, deployed on KDDI and OCI, and using the standards that Nexar, Cisco and UPC have been developing in the Internet Engineering Task Force (IETF). Running on OCI, the app relies on fast, flexible, and affordable compute capacity through OCI Compute and Oracle Container Engine for Kubernetes to run live data points for the AI algorithm.

"We are honored to be a part of this state of the art PoC demonstration with Nexar and Oracle in order to envision very advanced business use cases and their technical solutions developed by AECC,"  said Dr. Tomohiro Otani, General Manager of KDDI Corporation. "Such business use cases and edge/cloud computing technologies in the mobility domain are one of the most significant KDDI's strategies over 5G mobile networks. For that purpose, we are actively working for AECC as a member company as well as a chair of the PoC committee, and making every effort to contribute to a smart society in near future."

"Smart mobility services allow drivers to find available street parking efficiently and fast," said Toshimitsu Misawa, member of the board, CEO and president of Oracle Corporation Japan. "These services collect a very large amount of data which need to be processed near real-time. Oracle Cloud Infrastructure is designed to run nearly all applications more securely with enhanced data processing capabilities to increase speed and accuracy. We are glad to collaborate with Nexar and KDDI to continue to make advances in automotive edge services.

"In the near future, all vehicles will come equipped with dash cams, but the issue of how to upload the collected data without crashing the infrastructure will remain," said Bruno Fernandez Ruiz, CTO and co-founder of Nexar. "Partnering with AECC, KDDI and Oracle, we've discovered how to share this data scalably and economically in order to produce a collective value, enabling new applications such as parking detection. To deliver on the vision of real-time parking data, automakers will need to learn how to share data rather than assume that a siloed approach can succeed. With the AECC, KDDI and Oracle, we have accomplished just that."

Source: Nexar media announcement