The just better DATA research project is making a cross-partner dataset from a VDALI research project available to the public for the first time. The dataset is primarily intended for validating AI-based perception systems in automated and connected driving and thus makes an important contribution to an open and cooperative AI ecosystem in the automotive industry. The focus is on rare, safety-critical, and previously underrepresented traffic situations – precisely those scenarios that are crucial for robust AI applications in the mobility sector.

The hybrid dataset combines real-world and synthetically generated data and encompasses various sensor modalities such as camera, radar, and LiDAR. Supplemented by annotations and pseudo-labels, jbDATA thus creates a practical foundation for further developing, testing, and comparing AI methods under challenging conditions.

EICT has actively supported the development of the dataset from the outset, contributing its expertise particularly in legal, organizational, and infrastructural matters. In doing so, EICT has played a key role in ensuring that, for the first time, a cross-partner dataset of this kind could be successfully built and published.

Click here for the dataset: just better DATA – Dataset