# A Low-Complexity Hybrid Handover Strategy for LEO NTN: Balancing Stability and Link Quality

**Authors:** Khalid Aldubaikhy

PMC · DOI: 10.3390/s26051449 · Sensors (Basel, Switzerland) · 2026-02-26

## TL;DR

This paper introduces a new handover strategy for LEO satellites that reduces frequent switching while maintaining good connection quality.

## Contribution

A low-complexity hybrid handover algorithm that balances link stability and quality using a novel utility function and logistic-decay bonus mechanism.

## Key findings

- The HHS reduces handover frequency by 64% compared to SINR-based methods.
- The algorithm maintains 90.2% service availability with lower computational overhead.
- Validation was done using real-world Starlink TLE data in a high-fidelity simulator.

## Abstract

The deployment of low Earth orbit (LEO) satellite mega-constellations enables global broadband access, but their high orbital velocity demands frequent handover decisions that critically impact service continuity. Conventional strategies that maximize instantaneous signal quality often trigger excessive handovers, while stability-focused approaches may sacrifice link performance. In this paper, we propose the Hybrid Handover Strategy (HHS), a low-complexity algorithm that addresses this trade-off. The HHS utilizes a multi-attribute utility function that integrates the signal-to-interference-plus-noise ratio (SINR), satellite elevation angle, and network load with a novel logistic-decay stability bonus mechanism. We provide a formal mathematical analysis of the algorithm’s stability and performance trade-offs. To ensure industrial relevance, the strategy is validated using a high-fidelity simulator driven by real-world two-line element (TLE) data from the Starlink constellation. Results demonstrate that the HHS reduces the handover frequency by 64% compared to SINR-based benchmarks while maintaining service availability of 90.2%. The proposed algorithm delivers these improvements with significantly smaller computational overhead than machine learning approaches, making it suitable for resource-constrained on-board processing and ground terminals.

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/PMC12987173/full.md

## Figures

12 figures with captions in the complete paper: https://tomesphere.com/paper/PMC12987173/full.md

## References

36 references — full list in the complete paper: https://tomesphere.com/paper/PMC12987173/full.md

---
Source: https://tomesphere.com/paper/PMC12987173