Original title: AI-based Interference Mitigation for Power-domain Spectrum Sharing among LEO Satellites
Bowen Zhang, Barry Evans, Pei Xiao · September 2026
Low-earth orbit satellites bring high-speed internet to deserts, oceans, and rural places. They send data down to Earth using the Ku radio band. But when satellites from different operators line up, their overlapping radio signals clash. This problem is called co-channel interference, and it will worsen as more satellites launch. Researchers created a new system to stop this clash using power control. Satellites from one operator adjust how much power they transmit. This creates a clear power difference between the clashing signals when they reach ground users. User devices can then clean up the mess using successive interference cancellation or smart AI tools, which need a smaller power difference. Tests using Starlink and EutelsatOneweb signals show this greatly cuts interference and boosts spectrum efficiency.
You are at a noisy party trying to hear one friend. If two people talk at the same volume, you hear garbled noise. But if one speaks much louder, you can focus on that voice and ignore the quieter one. In this paper, satellites adjust their power so one signal is much stronger, and the receiver can pick it out first, then clean up the weaker one.
As more companies launch LEO satellite constellations, interference threatens service quality. This system allows multiple operators to share the same frequency bands by carefully adjusting power levels, increasing overall data capacity without needing new spectrum. Users in remote areas benefit from faster, more reliable internet as satellite networks work together.
The abstract does not discuss scaling to more than two satellite operators or handling dynamic satellite movements. It only evaluates with Starlink and EutelsatOneweb waveforms, so generalizability to other constellations is unknown. The method requires precise power control, which may be challenging in practice.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
AI-based Interference Mitigation for Power-domain Spectrum Sharing among LEO Satellites — Bowen Zhang, Barry Evans, Pei Xiao
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.