Collection and Validation of AI Datasets and Edge-enabled SNS Experimentation for 6G

Ubiwhere has a strong track record in the collection, generation, assessment and validation of datasets to support the training of AI models for next-generation digital services and 6G networks. This expertise builds on projects such as BeOpen, where Ubiwhere developed tools and methodologies to evaluate EU Public Sector High Value Datasets, addressing data quality, interoperability, representativeness and governance as key requirements for trustworthy AI. Ubiwhere covers the full dataset lifecycle, including real-time data collection, data cleaning, harmonisation and augmentation across heterogeneous sources, while accounting for different data governance and privacy requirements. The company is also experienced in defining validation criteria aligned with AI model training and AI-as-a-Service (AIaaS), positioning it strongly for STREAM-B-01 through the delivery of high-quality, reusable datasets for training, benchmarking and validation of AI models for 6G networks and AIaaS frameworks.

Complementing this data-centric expertise, Ubiwhere operates a proprietary Playground offering Open Labs for IoT, networking and AI experimentation. Equipped with diverse sensors, devices and a validated software stack, the Playground enables real-time data collection and the deployment of AI models across cloud and edge environments. Multiple interconnected edge nodes, linked to Ubiwhere’s datacenter, support the deployment, integration and validation of complex city-scale and network-centric use cases under realistic conditions and at high TRLs. This infrastructure is well aligned with STREAM-C-01, providing a flexible SNS experimental environment for end-to-end testing and validation.

Together, Ubiwhere’s dataset expertise and edge-enabled experimental infrastructure form a coherent foundation to support both data-driven AI training for 6G and large-scale experimentation of Smart Networks and Services in real-world contexts.