NavInfo Deployed AIRI AI Development for Autonomous Driving and Smart Cities

Pure Storage, Inc. announced that NavInfo (Europe) BV, an European R&D labs specialized in solutions for autonomous driving, simulations and advanced AI solutions, has deployed the integrated AIRI: AI Ready infrastructure based on NVIDIA DGX-1 and Pure Storage FlashBlade, an unified fast file and object storage platform.

NavInfo Europe incorporates the Advanced Research Lab for NavInfo Co., Ltd., and provides efficient simulation solutions which aim to promote safety and security for the automotive industry.

It needed to run complex AI models to produce accurate, up-to-date navigation maps. Vehicle manufacturers around the world rely on these maps for navigation and for autonomous driving, so it is essential for the company to be able to process very large data sets in real-time, 24/7. Accurately interpreting camera images for autonomous driving and keeping costs under control were two major challenges. it selected AIRI as it provides the performance infrastructure that is capable of real-time analysis of complex data sets and to train AI.

“Using FlashBlade and AIRI, we are able to provide clients in the automotive industry and in other domains (such as smart cities and asset management) with advanced and customized solutions“, says Frank Geujen, program management office director, advanced research lab, NavInfo. “The NVIDIA DGX processors in AIRI support highly complex calculations, but that capability must be supported by a storage system that can read and write large volumes of data very quickly. Pure Storage’s data centric design takes away many of the headaches we faced in terms of infrastructure and potential bottlenecks.“

Requirements for storage
For autonomous driving, the entire environment around the vehicle must be recognized, interpreted, and included in navigation, preferably in real time. NavInfo uses AI and deep learning to increase the quality of its products and services and uses the data from the many sensors that are part of every modern car. By using this data to train AI models, accuracy is increased.

“When designing and implementing an effective AI infrastructure, organizations often face challenges in integrating the various components,” says Theo van Teylingen, MD Netherlands, Pure. “It was clear that NavInfo needed fast time to insight: real-time analysis is essential to providing customers with highly accurate map information and navigation systems. In quickly and efficiently bringing together and analysing large disparate data sets, AIRI enables accelerated performance and provides peace of mind. In addition, it helps NavInfo keep costs under control, which has resulted in a high ROI.“

The models and large datasets used by NavInfo Europe place very high demands on storage. Analytics tasks must be able to run on different servers and sometimes even need to be moved between servers during the analysis process. Pure’s technology keeps the entire AI system performing smoothly. The next step is to provide AI-as-a-Service to automotive customers and customers from other sectors around the world. With this in mind, NavInfo is looking to expand its current AIRI system.

NavInfo AIRI case study