Telecoms AI synthetic datasets for Energy Optimisation in Stadium private 5G networks
Published by Digital Catapult
About this dataset
Funded by Innovate UK through the Assured 5G Technical Authority (A5G-TA) programme, Digital Catapult is building a sovereign collection of high-quality, openly accessible telecoms datasets. These datasets will support AI/ML development, network optimisation and, in the future, AI assurance for telecoms. The datasets provide realistic, diverse and repeatable private 5G network data that can be used to develop, train, validate and benchmark intelligent RAN optimisation solutions under representative network conditions. The 5G RAN energy optimisation use case focuses on understanding and reducing energy consumption while maintaining network performance, coverage, reliability and quality of service. The stadium synthetic dataset is generated using a simulated private 5G network environment, representing an eMBB-focused private 5G deployment with 10 cells and 500 UEs, capturing the highly dynamic network conditions associated with large-scale events. The 50-day simulation uses separate weekday and weekend profiles, with very low activity during non-event periods and increasing demand as events approach. Weekday traffic peaks during the 18:00–22:00 evening event period, while weekends show higher and more sustained activity from the morning through the evening. Elevated activity continues after peak periods as users depart the venue. This event-driven variation provides representative data for evaluating RAN resource utilisation and energy optimisation under rapidly changing network demand and user density.