Smart Energy Efficient Non-Terrestrial Networks: Towards 6G

The concept of Non-Terrestrial Networks (NTNs) as a component of 5G mobile networks was initially studied by the 3GPP during the work on Release 15, was codified in Release 17, and is expected to feature prominently in 5G Advanced (5GA) and 6G releases. AI/ML techniques play a crucial role in ensuring that 5GA and 6G networks will be energy efficient. Thus, there is a pressing need to explore how AI/ML techniques will impact the energy efficiency in ground systems, unmanned aerial systems, and satellite systems as constituent components of NTNs. Moreover, real-time Edge AI applications will impose a significant burden on the NTN infrastructure, including from an energy demand perspective. The Smart Energy Efficient NTN for 6G (SMARTEN6G) project is developing a methodology that will be utilized to emulate 5G, 5GA and 6G NTNs for planning and monitoring the energy efficiency of AI/ML techniques in transmission and service delivery chains, and to manage the energy demands of Edge AI applications. SMARTEN6G will employ a Digital Twin to emulate NTNs, especially where the layers are challenging to interface with for R&D purposes, such as geostationary and non-geostationary satellite systems and high-altitude platforms. The paper describes the main architectural choices adopted in the project to design a Digital Twin for the emulation of the NTN environment in the Open RAN (O-RAN) context, and for its integration in a cyberphysical system to construct a Proof-of-Concept prototype for the experimental evaluation of energy-aware AI/ML solutions.