Artificial intelligence (AI) has become a critical and indispensable component of vehicle software. Neural networks solve complex environment perception problems for driving functions. With this in mind, dSPACE, as an experienced solution provider for data-driven development, is already using AI in its products and to deliver new features.

The new dSPACE sensor vehicle is hitting German roads constantly, letting us use the collected environment data in combination with artificial intelligence to further develop dSPACE products. Our sensor vehicle is equipped with a rooftop box containing sensors (cameras, lidar, radar) and the powerful dSPACE AUTERA data logging hardware. This way, we flexibly collect realistic, specialized, and high-quality environment data for new dSPACE product developments, for customer projects, solution demonstrations, and for AI training and validation purposes. RTMaps is used as data logging software together with the live tagging application RTag. We use the Intempora Validation Suite (IVS) for data management and data selection.

Use Cases

Our top priority is implementing challenging simulation scenarios, such as driving under bad weather conditions. Here, the measurement data we collect helps dSPACE products and our customers' projects meet even the highest quality standards.

The recorded data from the dSPACE sensor vehicle also supports further important use cases within the field of data-driven development.

Data selection - developing AI models for automated tagging of unlabeled measurement data. This allows targeted searches in unlabeled data for objects such as cyclists and people, or for maneuvers such as lane changes and overtaking.

Data logging - Further development of smart logging algorithms and AUTERA hardware for smart logging only of data classified as "relevant".

Comparison of recorded and synthetic sensor data

Use of recorded measurement data for scenario generation