THESIS PROJECT
Station Keeping with Artificial Intelligence in GEO

Project Description

About

At Vinterstellar, we are developing next-generation mission analysis tools to make satellite operations more efficient, reliable, and innovative. Station-keeping (SK) is one of the most critical functions for geostationary satellites, ensuring orbit stability and mission success. While classical methods rely on deterministic equations or optimization-based schemes, we are now exploring how Machine Learning (ML) and Reinforcement Learning (RL) can open up new frontiers in maneuver computation.

Thesis Objective

The aim of this project is to design, implement, and benchmark an AI-based station-keeping maneuver computation algorithm within Vinterstellar’s simulation environment. You will explore whether ML (data-driven prediction) or RL (simulation-trained maneuvering policies) can provide faster, more robust, and more adaptable strategies than traditional methods.

What You’ll Gain

  • Hands-on experience with AI applications in orbital mechanics.

  • Insights into both classical and next-gen SK strategies.

  • Skills in ML/RL model design, simulation environments, and system benchmarking.

  • A chance to contribute directly to Vinterstellar’s mission analysis toolkit for future clients and missions.

You will work on:

  • Algorithm Development: Choose between ML or RL approaches to generate station-keeping maneuvers.

  • Simulation Setup: Configure Vinterstellar’s in-house SK simulation tools and datasets for training.

  • Two-Step Framework:

    1. Orbital maneuver computation (north-south and east-west control).

    2. Extension to angular momentum management for combined orbit and attitude planning.

  • Benchmarking: Compare AI-driven methods against classical solvers in terms of ΔV efficiency, computation time, robustness, and propulsion adaptability.

  • Integration: Deliver a prototype module ready for inclusion in Vinterstellar’s broader mission analysis toolset.

Location & Collaboration

  • Work closely with the Vinterstellar engineering team.

  • Duration: ~5 months (thesis project).

  • Flexible setup (remote or hybrid with touchpoints in Sweden).

How to Apply

Send your CV, transcript, and a short motivation letter to: career@vinterstellar.se

Applications reviewed continuously

Who Should Apply?

We’re looking for motivated master’s students in:

  • Aerospace Engineering (with focus on astrodynamics or guidance & control).

  • Computer Science / Data Science (with interest in space systems).

  • Engineering Physics, Applied Mathematics, or related programs.

Experience with Python, MATLAB, and a passion for combining AI with space engineering is highly valued.