If you are a climate tech agency dealing with massive data costs and slow response times for disaster alerts — this project developed an onboard image processing system that filters out low-value data. This ensures only critical information like fire detection is sent to ground stations, improving decision-making speed.
Hardware-Independent AI Processing for Earth Observation Satellites
Imagine a satellite that can think for itself instead of just taking photos and sending them back to Earth. Instead of wasting energy sending thousands of useless images of clouds, it only sends the important bits, like a fire or a plastic patch in the ocean. It uses a super-slim software layer that works on any computer chip, making it easy to update the satellite's brain while it's still in orbit.
What needed solving
Satellites collect more data than they can send to Earth, causing high latency and wasted energy. Current software is often tied to specific hardware, making updates difficult and expensive.
What was built
A Development Kit for lightweight applications, a Management Unit for onboard resource allocation, and a Simulation Platform for digital twin testing.
Who needs this
Who can put this to work
If you are a satellite manufacturer dealing with rigid software that only works on specific hardware — this project developed a hardware-agnostic orchestration tool. This allows you to offer a flexible platform where customers can deploy their own services regardless of the underlying chip.
If you are a surveillance operator dealing with high latency in facility monitoring — this project developed a secure unikernel-based deployment kit. This enables real-time processing at the edge, reducing the time between image acquisition and actionable intelligence.
Quick answers
How much does the system cost to implement?
Based on available project data, specific pricing or implementation costs are not provided.
Can this be scaled to other types of aerial vehicles?
Yes, the system is designed for flexibility and can be integrated into platforms ranging from high-altitude airships to drones.
What is the IP or licensing model for the software?
The project provides an open-source solution to democratize space access for manufacturers and developers worldwide.
How does this affect regulatory compliance for data security?
The system uses a secure-by-design unikernel architecture with data separation and cryptography to ensure robust and reliable operations.
How is the software integrated into existing satellite hardware?
It uses a hardware-agnostic approach and a Management Unit that allocates resources and adapts priorities dynamically across missions.
Who built it
The consortium is heavily industry-driven with an 80% industry ratio, comprising 5 partners across 3 countries (France, Poland, Romania). Led by Thales Alenia Space, the group includes 2 SMEs, indicating a strong focus on commercial viability and industrial application rather than purely academic research.
Contact Thales Alenia Space France SAS
Talk to the team behind this work.
Contact us to explore licensing the open-source unikernel tools for your edge infrastructure.