handoff at seventeen thousand miles per hour

Problem statement

No standard exists for seamless handover between LEO satellite cells for broadband service. Terrestrial cellular handover protocols assume slow-moving users, predictable cell boundaries, and sub-5ms latency. LEO satellites move at ~7.5 km/s, a single satellite's coverage may last only seconds to minutes, round-trip latency is 20-40ms, and Doppler shift from satellite motion (~48 kHz at Ka-band) exceeds the tolerance of existing synchronization protocols. The handover problem is fundamentally different from any terrestrial mobility scenario.

Why this matters

SpaceX Starlink, Amazon Kuiper, OneWeb, and others are deploying or planning megaconstellations totaling tens of thousands of LEO satellites for global broadband. Without reliable handover, these networks cannot deliver continuous connectivity for maritime communications, aviation broadband, rural 5G extension, or disaster response. Dropped connections during handover are already a recognized pain point in current Starlink service. The market for satellite broadband is projected to serve hundreds of millions of unconnected people.

What’s been tried and why it hasn’t worked

3GPP Release 17-18 added NTN (Non-Terrestrial Network) support for 5G, but handover mechanisms are adapted from terrestrial protocols with timing adjustments rather than designed from first principles for orbital dynamics. Three fundamental problems remain unsolved: (1) predictive handover requires accurate orbital mechanics computation at the terminal, but consumer devices lack this capability and real orbits deviate from predictions due to atmospheric drag and solar radiation; (2) multi-layer constellation handover (between satellites at different altitudes, e.g., Starlink shells at 340-570 km) has no protocol; (3) security context transfer during high-frequency handovers has no validated mechanism — pre-authentication and trust transfer must happen potentially every few seconds.

What would unlock progress

An orbital-aware handover protocol that uses satellite ephemeris data to predict handovers before they occur, combined with make-before-break beam switching that maintains connectivity during transitions. The key unsolved piece is the multi-layer case — when a terminal should switch from a low-altitude satellite (lower latency, smaller coverage) to a higher-altitude one (longer coverage, higher latency) requires a cost function that balances latency, throughput, and handover frequency in real time.

Entry points for student teams

A team could simulate LEO constellation handover performance using open tools (ns-3 NTN module, STK) to quantify how handover failure rate varies with constellation parameters (altitude, inclination, inter-satellite links). Alternatively, a team could design and evaluate a predictive handover algorithm using TLE orbital data. Relevant skills: telecommunications, orbital mechanics, network simulation.