SIBILA AI inference engine
As part of the Spanish Recovery, Transformation, and Resilience Plan, Celeres (formerly GTD) delivered the SIBILA initiative. This innovative on-board Artificial Intelligence solution optimises data transmission from satellites to Earth. Built entirely on Edge Computing technology, the system reduces latency and significantly enhances threat detection by processing data directly on the spacecraft.
Funded by the Spanish Ministry of Defence through CDTI and DIGEID, the project was executed with the strategic backing of Airbus DS GEO and Airbus Crisa. The development of SIBILA built directly upon our proven engineering heritage with the VPDU-BOX-AI.
Our team successfully delivered an architecture that embeds machine-learning algorithms to cut data downlinks by up to 80%. Celeres led the design of specific AI models for on-board anomaly detection, data prioritization, and intelligent compression. We engineered the software prototype to leverage robust inference engines specifically optimized for radiation-hardened space processors. To guarantee adaptability, we integrated continuous delivery pipelines that support in-orbit updates. This allows the adaptive models to be modified during flight. By processing raw sensor and telemetry data in space, SIBILA ensures that only high-value insights are transmitted to ground stations.
Explore more missions
Other programmes where our ground systems are in the operational loop.
Security
Ongoing
SECPRO: French Ministry of the Armed Forces and Veterans Affairs
Celeres integrates physical and digital security solutions to protect French military infrastructure, from design through operational deployment.
Ground
Ongoing
SALTO (Reusable Launcher Technologies)
Within Horizon Europe's SALTO project, Celeres develops the Ground Autonomous Flight Termination System (GAFTS) and validates SAMMBA launch-complex services to mature Europe's reusable launcher operations to TRL6.
Have a project in mind, let's talk
