ESA SHINE Project Smart Handover for Integrated 5G Non-Terrestrial Networks

Integrating satellite non-terrestrial (NTN) and terrestrial (TN) networks enables ubiquitous network access, requiring 5G user equipment (UE), such as handheld devices or vehicle devices, to connect to both networks directly. In this scenario, UEs may need to perform handovers between terrestrial and non=terrestrial networks. Optimizing the number of handovers and the success rate of switching between networks is critical to minimize service interruption. Conventionally, handover can be triggered by signal strength measurements, location information, and other parameters such as elevation angle, Timing Advance (TA) values or Doppler values exceeding a predefined threshold. However, the handover decision is also impacted by the high mobility of UEs and satellites, strict Quality of Service (QoS) requirements such as reliability and latency, and the Quality of Experience (QoE) in terms of UE service satisfaction. Additionally, the different architecture between the non-terrestrial and terrestrial networks, the changing communication environment, and the end-user requirements may lead the conventional handover to make late/early decisions and switch to the wrong network/cell. For this, ML-aided handover decisions can be a solution for the TN-NTN integrated network to deal with the network's partial information, uncertainty, rapidly changing environment, and strict time constraints.