jotne_4: LLM-Assisted Automated Flight Operations Procedure Generation and Validation for CENSSAT-1

AIM: To design, implement, and validate an LLM-assisted Operations Preparation Assistant (OPOS) framework that automates the generation and verification of spacecraft Flight Operations Procedures (FOPs) for CENSSAT-1. 

Bildet kan inneholde: astronomisk objekt, teknologi, det ytre rommet, satellitt, telekommunikasjonsteknikk.

Required Data & Inputs

  • Mission & User Manuals: CENSSAT-1 technical documentation, operational constraints, subsystem descriptions, and equipment specifications. 
  • Flight Operations Procedures (FOPs): Industry standard operational procedures and historical flight sequences. 
  • Spacecraft Databases: Mission-specific databases containing Telecommands (TC), Telemetry (TM), parameters, definitions, and related operational constraints. 

Objectives

  • Knowledge Integration & Retrieval: Build mission-specific knowledge loaders and graph databases (RDF conversion) for CENSSAT-1 to support relationship-based queries across spacecraft components, subsystems, TCs, and TMs. 
  • Automated Procedure Generation: Develop and fine-tune AI agents to converse with operators and dynamically generate step-by-step Flight Operations Procedures (FOPs). 
  • Validation & Consistency Checking: Implement automated verification mechanisms to cross-check AI-generated procedures against official spacecraft databases, operational constraints, and historical FOP standards. 
  • Deployment & Integration: Integrate CENSSAT-specific knowledge loaders, query tools, and AI agent enhancements into the standard architecture.. 

 

This thesis project is supervised by the CENSSS partner Jotne Connect. 

Contact persons are Kjell.Bengtsson@jotne.com and aditya.gupta@jotne.com

Tags: LLM, AI, satellite modelling
Published Nov. 4, 2026 11:29 AM - Last modified Nov. 4, 2026 11:29 AM

Supervisor(s)

Scope (credits)

60