tech
A satellite just learned to find things on its own — here's what that means
In April, for the first time ever, an Earth observation satellite found what it was looking for, all on its own.
TL;DR
- An Earth observation satellite, Yam-9, autonomously found targets using an onboard vision-language model (VLM) for the first time.
- The VLM, Google DeepMind's Gemma 3, is designed for edge applications, running on limited hardware in space.
- This AI capability could make space sensors more useful by performing initial data triage on orbit, reducing data volume for analysts.
- The technology enables 'always-on' orbital monitoring and interaction with satellites, such as specific border surveillance.
- Loft Orbital, the spacecraft provider, operates on an infrastructure-as-a-service model, with Yam-9 serving as a pathfinder for orbital AI projects.
- NASA JPL developed the NAVI-Orbital software package to integrate the Gemma 3 VLM onto the satellite.
- Other companies like Planet Labs are also exploring AI applications, including VLMs, on their satellites.
- Future applications could include digital assistants for astronauts exploring other planets.