6G-SANDBOX

First, the validation of the use case and energy-saving strategy demonstrates how AI-driven and O-RAN-based programmability can help MNOs reduce CO₂ emissions and associated OPEX, contributing to societal and environmental sustainability in line with global SDG goals.

For instance, results show that, for an average throughput of 15 Mbps per UE and a maximum outage probability of 5%, the targeted cells could be switched off for approximately 40% of a week. When considering 5 Mbps and 10% outage, this value could increase to around 70%. These policies, managing the energy-QoS trade-off could be defined by the operator and considered by our developed solution during inference.

Second, we have advanced the implementation of our O-RAN framework (AI Engine plus RICs), developing and validating its compatibility with RAN emulators such as the Keysight RIC Tester. This progress will increase the TRL of the infrastructure by demonstrating scalability and conformance with O-RAN standards, while enabling the development of novel and innovative use cases in future projects.