Multi-Criteria Conditional Handover in 5G NTN Mega-Constellations

Mobility management in 5G Non-Terrestrial Networks (NTN) must address the predictable yet continuously evolving connectivity conditions of large Low-Earth-Orbit (LEO) mega-constellations. In such environments, multiple satellites may simultaneously provide feasible access while the selected attachment point determines the subsequent end-to-end transport path towards gateway infrastructure. Consequently, mobility decisions driven solely by signal-strength indicators do not necessarily ensure either long connection persistence or transport-efficient routing, since they do not account for constellation-level dynamics. This paper proposes a multi-criteria Conditional Handover (CHO) framework enabling predictive ordering of candidate satellites based on combined maximal residence-time potential and anticipated transport efficiency. Predictive indicators are derived from a constellation digital twin implemented using the Fraunhofer FOKUS OpenLANES toolkit, providing awareness of routing evolution beyond instantaneous visibility conditions. Evaluation on a large LEO mega-constellation shows that transport-aware attachment improves latency consistency, while persistence-aware selection reduces unnecessary handover activity. The results highlight the benefit of integrating network-level performance indicators into CHO design for future NTN deployments.