Despite the restrictions imposed by Covid-19, the project consortium of KI Data Tooling succeeded in reaching the project’s first milestone on time. With that, the consortium created the foundation for the data generation processes, as well as the development of methods in the project.
The goal of KI Data Tooling is to develop a complete data solution for the training and validation of AI functions for automated driving. The project aims to create a comprehensive database of real data recordings, synthetically generated data, and synthetically augmented real data recordings. The data set will encompass different sensor modalities, such as, video, Lidar, and Radar. Additionally, a set of methods and tools will be developed to make preparation, storage, and processing of data more efficient, aiming for automation whenever possible.
Reaching the first milestone means that the foundations for the two main data sources (real data and synthetic data) have been laid and the necessary preliminary work has been completed. Next, data generation will begin. First sequences of real data recordings are already available in the project. All the necessary foundations to start the synthetic data generation have been laid. Additionally, work on methods, e.g., corner case detection, compression, and context, has started and has already led to first publications from the project.