Challenge your automated driving Software & accelerate your Development!

Test your ADAS/AD systems in a digital environment based on real data.
Illustration zur vereinfachten Darstellung der Überschrift

Multiply your current test case coverage

Illustration zur vereinfachten Darstellung der Überschrift

Detailed analysis of scenarios

Illustration zur vereinfachten Darstellung der Überschrift

Verify your Level 2+ to Level 5 systems

Illustration zur vereinfachten Darstellung der Überschrift

Reproducibility
of the test cases

Simulation platform for testing & validating systems for automated driving

The different phases in the development and testing of AD/ADAS functions and algorithms for automated driving require extensive resources. With AAI ReplicaR, you can find the real world in a virtual environment, with all the advantages it offers. From efficient deterministic maneuver tests up to randomly generated scenarios in virtual endurance runs, AAI ReplicaR supports you in all stages of your development.

Virtual endurance run -
realistic test drive simulation

With our naturalistic traffic simulation, it is possible to generate an infinite number of realistic scenarios mechanically and thus within a very short time by using AI. In the form of virtual endurance runs, a high test case coverage can be achieved with little manual effort, making your final product more seamlessly tested. Through automated analysis and extraction, unforeseen borderline cases (so-called corner cases) are detected, which offers you a fast evaluation and the possibility to use these situations as scenarios for further optimization and testing of your systems.

Realistic digital environment

Our simulation environment is built from real data, according to "create a replica of the world". The combination of a 3D environment (static simulation) and the simulation of traffic events (dynamic simulation) enables a realistic virtual test ground for your ADAS/AD systems. We create maps in OpenDRIVE format based on real road geometries from various sources. You can procedurally generate your simulation environment based on OpenDRIVE data at the push of a button and extend it with available assets from our database. If you require a one-to-one copy of a specific environment, our 3D artists will be happy to create highly accurate reproductions/digital twins for you. By storing meta information, such as material properties of the 3D objects, the simulation environment is enhanced.

FAQ Section

Can I integrate individual modules, features or service results into my existing simulation environment or, vice versa, transfer them into AAI ReplicaR?

Yes, AAI ReplicaR is a modular platform for data or knowledge based approaches. Each module is developed considering industry standards as an input or output generator. Customers can use either a single module, such as the Map Editor, Sensor Simulator, Scene Generator (3D), or Scenario Generator (Traffic Module), or the complete toolchain. Also, third-party maps, for example, can be used as long as they meet industry and quality standards. The ReplicaR approach allows the highest flexibility for the end user. Any existing toolchain, either for synchronous or asynchronous mode, or any XiL application, can be connected.

How do you prove that the "true-to-life environment" in ReplicaR really corresponds to reality? How do you validate your features?

Automotive Artificial Intelligence follows the approach: verification before validation, to ensure that our toolchain meets our customers' requirements. As an example, for AAI's naturalistic traffic, we scientifically compare our data to real world data. Our approaches to verification are purely scientific and contractually transparent to the customer.

How can I integrate my own ADAS/AD algorithms into RelpicaR? (especially the algorithms for sensors, sensor fusion, perception and driving function/action)

AAI ReplicaR provides an SDK with all interfaces, either based on object lists or raw data, in synchronous mode or in asynchronous mode, etc. We are open to ego algorithms for automated driving functions and action. All data queries are provided via OSI (Open Simulation Interface) if required.