AI powered mapping and localization for autonomy


accurate | robust | scalable | adaptable

Learn more

Mapping and localization are critical for autonomous driving. YDRIVE offers a robust yet cost-effective solution using an innovative combination of geometric computer vision and deep learning methods.


Robust

High persistence in dynamic urban conditions. Resilient to adverse weather conditions including snow and rain.

LIDAR-less

Uses just cameras that are already present. Easily adapted to perform with customers camera choices.

Cost Effective

Edge processed, vision based system, with extensive pipeline automation results in industry leading cost savings.

Scalable

Camera based solution allows for large scale crowd sourcing. Highly automated processing leads to quick map updates.

HD Mapping


Rapid and large scale HD maps at a fraction of the cost

HD maps provide detailed information necessary for safe and reliable operation of autonomous vehicles. Information such as road sign and lane level traffic light positions, pedestrian crossings and even right-of-way rules in complex intersections are provided by the HD map. Our HD maps provide industry-standard 10 cm accuracy, while our deep learning based mapping pipeline enables rapid scaling at the rate of 1000 km per day.


Real world interactive map samples
Four-lane road

Reconstruction of a straight four-lane road with semantics.

Narrow alleys

Reconstruction of narrow alleys and parking lots with semantics.

Localization


10 cm accurate all weather localization

To fully harness the power of HD maps, autonomous vehicles need to accurately pinpoint their position. Known positioning systems such as GPS do not provide the accuracy nor reliability required to safely operate autonomous vehicles. Our localization stack combines geometric computer vision and semantic understanding with INS and GNSS to provide accurate and reliable localization, with the ability to handle adverse weather conditions.

Simulation


Dynamic simulations based on accurate mapping data

Autonomous car development requires validation of many different software systems including numerous machine learning models in realistic driving situations. Our real world mapping data enables us to synthetically generate highly accurate and diverse worlds for autonomous driving simulations. Our technology allows for easy integration into diverse environments and is compatible with leading simulation engines.

Team


We are researchers and engineers passionate about computer vision and machine learning

Our founding team has over 30 man-years of experience at Fortune 50 companies building products at the intersection of state of the art research and cutting edge engineering. Our combined experiences include developing large scale software for Augmented Reality, Mapping, Drones, High availability cloud, Operating systems and Embedded systems.

Let's Get In Touch!


We would love to partner with you. Send us an email and we will get back to you as soon as possible!