Researchers have built an AI model that identifies individual satellite components in radar images and tracks them across a sequence, even when a satellite rotates and a part disappears from view. The work, involving the Alan Turing Institute and the University of Birmingham, could help improve space monitoring as orbit becomes more congested.
Watch Desk analysis
What happened
The model can pick out components including solar panels, thrusters, antennas and robotic arms in high-resolution radar imagery, according to UK Defence Journal’s report. Its temporal memory encoding uses earlier images to help interpret later ones, maintaining a track when a component is temporarily hidden by the satellite’s movement.
The imagery comes from a sub-terahertz inverse synthetic aperture radar system developed by Birmingham’s Pervasive Sensing Group. The research, led by the Turing’s Defence AI Research Centre alongside the university, was presented at the European Radar Conference 2026.
Why it matters
Satellites and other objects in orbit are currently monitored manually, the report says, while growing congestion raises collision risks. Identifying and tracking individual parts from space could give operators a more detailed picture of what an object is doing. The researchers also say the capability could support satellite docking and rendezvous, which are relevant to maintenance and debris removal.
This is a research development, not evidence that the model is already operating on satellites or has improved collision avoidance. The distinction between a promising tracking capability and a working orbital service is, inconveniently, where the engineering lives.
Our read
The notable step is persistence: recognising a component again after it has slipped out of view. That is a useful capability for monitoring objects in motion, and a more specific advance than simply saying AI can “analyse satellite imagery”. The next test is how well it works across varied objects and real operating conditions.
What to watch
- Whether the team publishes evaluation results on tracking accuracy across different satellites and viewing conditions.
- Whether the system is tested with radar imagery collected in orbit.
- Whether researchers demonstrate a practical use in space-domain awareness, docking or debris removal.
Discussion spark: Should space agencies prioritise AI that tracks objects from orbit, or invest first in improving the radar and telescope systems already used to monitor space?
Sources and evidence
- Source update (12 October 2026, 04:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.