ICE2THRUST

Project Overview

ICE2THRUST represents a groundbreaking approach to space propulsion, developing the world's first In-Situ Resource Utilisation (ISRU) end-to-end process chain that transforms ice into thrust. This innovative system utilizes water as a propellant, decomposed by an electrolyser into hydrogen and oxygen for subsequent use in a thruster.

As Lead Research Scientist for RL-based autonomous navigation systems, I'm responsible for developing the intelligent control algorithms that enable autonomous proximity operations, docking, and propellant refilling procedures.

Technical Approach

🚀 Solar-Electric Water Electrolysis Propulsion

Development of high-efficiency propulsion systems using water as propellant, achieving the highest specific impulse of all storable chemical propulsion systems.

🤖 Autonomous Proximity Operations

RL-based algorithms for autonomous docking, proximity operations, and propellant refilling procedures with machine learning-enhanced decision making.

⛏️ In-Space Resource Utilization

Advanced systems for water extraction from icy regolith and integration with propulsion systems for self-sustainable space operations.

My Research Contributions

Simulation Framework Development

Modular spacecraft simulators: Designed and built modular 3DoF and 6DoF spacecraft dynamics simulators for proximity operations, covering a range of parametrizable satellite configurations and mission profiles.
Proximity operations scenarios: Developed simulation scenarios covering approach, station-keeping, docking, and propellant refilling operations under realistic disturbances and uncertainty.
Scalable environment design: Built the simulation infrastructure to support large-scale parallel training of deep RL policies using IsaacLab/Omniverse.

Autonomous Navigation and Control

Deep RL policies for rendezvous: Designed and trained deep RL policies for robust autonomous rendezvous and docking under uncertainty, including unknown mass properties and thruster noise.
Hardware-in-the-Loop validation: Integrated trained controllers into Zero-G floating-platform Hardware-in-the-Loop (HIL) setups to study and validate sim-to-real transfer.
Sim-to-real transfer: Developed and evaluated domain randomization and adaptation strategies to bridge the gap between simulation and the physical floating platform testbed.

Project Resources

Official Project Website Detailed Project Description EU CORDIS Entry

Solar-Electric Water Propulsion System