Validating sensors of autonomous vehicles in virtual drives

16th February 2022

dSPACE Ltd

Kiera Sowery

0 0

dSPACE is launching AURELION, a sensor simulation solution that provides high-resolution visualisation for testing and validating functions for autonomous driving.

The solution, which can be operated either in the cloud or locally, generates photorealistic images in real time for camera simulation and, using ray tracing, an exact environment for radar and lidar simulation. With AURELION, developers validate algorithms for autonomous driving by means of simulation during virtual test drives, long before a prototype hits the road.

AURELION can be used throughout all phases of the development process, for example, software-in-the-loop (SIL) testing, hardware-in-the-loop (HIL) testing, or simultaneous validation in the cloud. The solution supports the development of functionalities and training data based on artificial intelligence, including neural network training and testing.

A powerful 3-D rendering engine, high-precision dSPACE simulation models, and realistic 3-D assets such as pedestrians or vehicles enable accurate simulation of sensors, environments, weather conditions, light conditions (day, night), and materials. Developers can thus simulate a wide range of scenarios and test corner cases that very rarely occur in real test drives.

AURELION draws on an extensive library of sensor models that is always up to date. This means that new sensors can be replicated in simulation solutions long before they are launched on the market. To this end, dSPACE has entered into cooperation agreements with the world's leading sensor manufacturers and is continuously expanding its collaboration with developers of camera, lidar, and radar technologies. There is also the option of integrating third-party sensor models.

The solution bundles functionalities from dSPACE MotionDesk and SensorSim into one single product, and also provides extended functionalities and more precision in the field of sensor simulation. In combination with other dSPACE tools, such as the ASM simulation models, the VEOS simulation platform, and ModelDesk for parameterization, dSPACE provides anend-to-end solution for testing and validating the autonomous vehicle (AV) stack.