Imagine trying to run a powerful computer in a place where the air is filled with tiny, invisible bullets that break electronics. Most AI chips are too fragile for this, so they usually stay on Earth. This work creates a 'tough' version of these chips and the software to run them, allowing spacecraft to think and make decisions on their own without waiting for instructions from home.
Current AI hardware (GPUs) is too fragile for space because radiation destroys their high-density transistors. This forces space missions to process data on Earth, causing delays and limiting autonomy.
A prototype hardware platform using radiation-hardened processors and a matching software suite to run ML algorithms in space.
If you are a satellite manufacturer dealing with the lag of sending data back to Earth for processing — this project developed radiation-hardened computing platforms that allow AI to run in situ. This means your satellites can analyze data instantly in the harsh environment of space.
If you are a space surveillance firm dealing with hardware failure due to cosmic radiation — this project developed a prototype hardware using only space-ready components. This ensures AI algorithms for object detection remain functional in high-radiation zones.
If you are a probe developer dealing with limited bandwidth for remote control — this project developed a virtual environment and a suite of ML algorithms for space deployment. This enables autonomous navigation and scientific discovery without constant ground support.
Based on available project data, the specific commercial price is not listed, though the EU provided a contribution of EUR 1,499,831 for the research and development phase.
The project aims to reach TRL-4, which is a laboratory validation stage. It is currently a proof of principle and not yet ready for full industrial scale.
Based on available project data, specific licensing terms are not provided; however, the project involves a consortium of 6 partners including universities and industry.
The project period runs from 2023-10-01 to 2026-09-30, meaning the TRL-4 validation will be completed by late 2026.
The project develops specific software designed to run AI algorithms on the most advanced space-hardiness proven processors.
The consortium is heavily research-driven, consisting of 3 universities and 2 research institutes, with only 1 industry partner (17% industry ratio). This suggests the output is primarily technical and academic, though the inclusion of an SME indicates a path toward commercialization of the software and hardware prototypes.
Contact the Katholieke Universiteit Leuven research office regarding the ASAP project.
Contact us to bridge the gap between this TRL-4 prototype and your space hardware needs.
