Required Data & Inputs
- Historical and real-time Telemetry (TM) and Telecommand (TC) data logs.
- Anomaly Event Logs, and satellite operational manuals.
Objectives
- Develop an automated time-series pre-processing pipeline to ingest multi-channel telemetry into an LLM-compatible vector database or Knowledge Graph.
- Implement fine-tuned LLM architectures or RAG pipelines capable of analyzing the observed telemetry anomalies.
- Build an intelligent Telecommand (TC) verification and monitoring assistant that evaluates sent commands against current satellite health states to prevent hazardous operations.
- Validate the LLM framework against synthetic and real mission anomaly scenarios, ensuring efficient real-time execution capability on ground control .
This thesis project is supervised by the CENSSS partner Jotne Connect.
Contact persons are Kjell.Bengtsson@jotne.com and aditya.gupta@jotne.com