Handover Mechanisms for Integrated Terrestrial and LEO Satellite Networks: A 6G Survey and Research Roadmap
Year: 2026
Type: conference-paper
Abstract: The seamless integration of terrestrial networks (TN) and non-terrestrial networks (NTN) is imperative for the realisation of 6G ubiquitous connectivity. It is evident that traditional handover mechanisms, which have been designed for static 2D networks, are ineffective in dynamic 3D environments. The movement of LEO satellites at 7.5 km/s results in handovers driven by orbital mechanics rather than user mobility, and 20–40 ms propagation delays render measurement reports obsolete before processing. This paper provides a systematic survey of TN-NTN handover mechanisms based on analysis of recent peer-reviewed literature and 3GPP Rel. 17–19 specifications. The solutions are classified into four paradigms: measurement-based, location-aware predictive, AI/ML-enabled, and hybrid network-UE controlled. These solutions are evaluated under consistent 3D network assumptions. The findings of the present study demonstrate that predictive methods have the capacity to reduce latency by up to 21× in comparison with reactive baselines (from 2.1 s to 100 ms), whilst deep reinforcement learning achieves 92–96% success rates with 18–32% fewer handovers. However, it is notable that over 80% of AI/ML proposals necessitate onboard resources that exceed the capabilities of LEO satellites (<10 GFLOPS), thereby underscoring a conspicuous deployment gap. The identification of four research priorities is proposed: cross-domain unification, multi-objective optimization, context-awareness, and RAN architecture adaptation. These priorities are aligned with 3GPP Rel. 19/20 timelines. The necessity of addressing unification and experimental validation is paramount in order to realise the objective of seamless 3D mobility in 6G networks.
Authors: Ainur Daurembekova, Hans D. Schotten
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7226025285
