top of page

The New Standard
in Road Condition & Traffic Data Analysis

It processes 100 times more data than any other navigation app

Atlas Navi app helps drivers avoid traffic by using the smartphone’s camera and artificial intelligence to”see” and analyze the road constantly

The app detects vehicles (including police), pedestrians, traffic, roadwork, potholes and accidents and sends relevant information to other drivers, helping dissipate road congestion faster. It processes 100 times more data than any other navigation app and it’s constantly on the lookout for road changes to automatically share with the network of connected drivers thus routing traffic more effectively than ever before.

Data Center Performance Anywhere

Data storage depends on the amount of data collected and how it is stored.
The more data you have, the more you can train and learn.

A.I Supercomputing Stations

The World’s First AI System Built on NVIDIA  offering unprecedented compute density, performance, and flexibility in the world’s first 5 petaFLOPS AI system

Massive amounts of data

Each data-collection car generates upwards of two petabytes of data with billions of images per year. The data center uses inferencing to ingest, curate, and annotate this data for use in the AI model training.

Most Secure AI System

The most robust security posture for your AI enterprise, with a multi-layered approach that secures all major hardware and software components.

Optimized Workflow

Cloud instances to increase performance, remove bottlenecks, and optimize workflow.

 multi-tiered storage options that produce higher throughput and can serve millions of small, random files to GPU-based training servers accelerating overall training times.

Teaching Maps to identify and suggest alternative routes 

The app uses either the smartphone’s camera or any existing dash camera connected via USB to the smartphone.The app analyzes what the camera is capturing and scans the road for problems.All problems are detected and reported automatically or manually. Traffic congestions are avoided more effectively than with any other solution currently on the market.

We Take Pride in Our Numbers

12

Years of Experience

250M

Orders processed in our servers

80

Transportation apps developed

22

Countries World Wide

6M

Total app downloads

Vision & Achievements

Machine learning / computer vision neural networks that interpret the road as seen by smartphone cameras

We trained the neural networks on 100 TB of images and videos to identify various objects on the road that are of interest.


We encapsulated these into a mobile app is custom technology which was faster than the available solutions on the market in terms of processing the video.


We incorporated a mapping provider that would allow us to change routes based on specific segments of the roads which needed to be avoided.

Atlas Navi apps already built on Android and iOS with 90% of functionality available. We have been testing the apps for over 6 months internally to make sure they work. Being a real utility navigation app relying on computer vision (A.I.) analysis of the video feeds from the smartphone cameras, we had to make sure the huge data generated by the real time detections of traffic, potholes, accidents, road construction, available parking spaces (and others) work perfectly.

Our Partners

Nvidia Accelerator

Waymo
Dataset

HERE
Maps

E.U
Funding

Dive Into the Notable technology overview

Hardware infrastructure

 Cutting-edge A.I. servers, making us the leader in A.I. infrastructure in our home country of Romania. It includes hardware capable of training terabytes of images and videos to improve detection engines over 75 times faster than conventional servers. It allowed us to “teach” our algorithms how to detect objects by using massive, ever-expanding datasets. With more data, the algorithms become constantly better, which leads to more users and, in return, more data for training and so on.

Massive datasets for training A.I. algorithms

Through our partnerships with NVIDIA and Google Waymo’s datasets of driving data, as well as our own, we have assembled a library of 100 TB of images with labelled vehicles, pedestrians, potholes, poles, accidents, road work, police vehicles and many more.

Deep Neural Network training:

Training our algorithms on this massive data volume has provided a very accurate detection engine, embedded in the mobile apps and updated regularly. The deep neural networks analyze in real time everything the camera sees

A.I. inference through our Object Detection engine

Our custom algorithms run a process called A.I. inference on local hardware, without the need for an internet connection. They store detections and upload reports as soon as server communication is possible.
The algorithms are able to extract, analyze, process and store video feed

Edge A.I. algorithms

The detections are made on the device, rather than on a centralised server. This decentralized way of analyzing data gives our network incredible power, flexibility and infinite scale. Our efficient software algorithms avoid draining the phone’s battery whilst analyzing videos over 30 times / second

Cloud Video Storage:

Road videos can be stored on device or in the cloud, shared or kept private. The videos serve as a historic video trip.
 

Integrated in-app web3 wallet

NFT licensing for 3D vehicles in the app.
Drive-to-Earn mechanism and variable reward system based on time, day, location and vehicle NFT

Fleet dispatching & Matrix Routing

Matrix Routing algorithms for multi-stop trips
Fleet dispatching, job allocation and asset tracking for fleet managers.

bottom of page