Automotive ADAS & AD simulation
Creating real-world scenes at scale
We offer 3D world twins as environment models for simulation of advanced driver assistance systems (ADAS) and autonomous driving (AD) systems. Reconstructed from real-world geodata with our AI and 3D technology.
Our 3D world twins are ready to be integrated and imported as static scene layers in existing simulation frameworks. We currently export in FBX and OBJ format, support a direct Unreal Engine integration, and offer different data integrations for MathWorks RoadRunner.
We support ASAM OpenDRIVE and every 3D world twin is generated in sync with an existing HD-map.
Our 3D world twin technology brings several advantages for your sensor and perception simulation:
With our rapid generation of 3D world twins we enable you to simulate at scale, reduce cost and accelerate time to market of your ADAS and AD solution.
Let’s define your area of interest and you will rapidly receive your 3D environment ready for simulation!
We offer turn-key 3D models as well as several data layers from our AI analytics solution. Let’s discuss your need!
Variations for billions of miles
Rapidly create infinite variations of the 3D scene
Our 3D technology allows us to generate infinite numbers of variations of a single scene.
We enable you to simulate your ADAS and AD stack in a large variety of synthetic environments which are rapidly generated.
This increases your synthetic data diversity, enables you to run brute force testing and ultimately increases your perception robustness.
We are here to support you towards a billion miles in simulation and testing – Unlocking the full potential of virtual vehicle development.
PBR Materials for sensor simulation
3D scenes optimized for physical sensor-realism
As our 3D world twins are object-based reconstructions, we are able to support physically based rendering (PBR) of materials for all objects in the scene.
We enable you to simulate your visual sensor systems at scale as all objects are reflecting and absorbing light waves based on realistic material properties.
We are here to support your simulation and testing of virtual camera, radar and lidar systems in our photo- and sensor-realistic 3D world twins.
Enabling Ground truth data
Accurate synthetic training data at scale
As our 3D world twins are object-based, semantic reconstructions, every object in the scene is classified and can thus be exported as ground truth information.
With our procedural 3D scene generation we enable you to rapidly create large amounts of synthetic training data for your perception algorithms.
We are here to support your AI development with perfect ground truth in e.g. semantic segmentation or depth maps.
Other Use Cases:
We are coming from automotive, but we are not limited to that!
Our 3D world twin technology can be applied in several other industries.
Rapidly creating 3D content from real-world location data can unlock the next level of your own use case? We are happy to support you with a 3D world twin, based on your area for your specific use case.
Get your sample data now!
Do you want to experience and explore our virtual 3D world twins in your pipeline?
Check out the sample data from our current beta stage!