SKYWALKER - Autonomous Control of a Free-Floating Space Manipulator in Simulated Microgravity Using Reinforcement Learning

Davidsson, Dadi Hrannar ; Svendsen, Glenn Gadensgaard ; Miguelez De Salas, Bruno

4. semester

Robotics, M.Sc.

2025

2025-06-04

83

This project explores the use of Proximal Policy Optimization (PPO) to enable autonomous control of a robotic manipulator mounted on a free-floating base in a simulated microgravity environment. A simplified setup of ESA’s Orbital Robotics Laboratory (ORL) was developed in Isaac Lab, including a robotic arm, fixed grasping points, and a frictionless floor. The system was trained using curriculum learning and evaluated through structured acceptance tests covering Point-to-Point motion, grasping, base relocation, and multi-step traversal. Results show that PPO can produce smooth and accurate behaviors without predefined trajectories in early tasks. However, the final traversal task was not solved, highlighting challenges in long-horizon planning. Still, the project demonstrates a functional RL-based control pipeline, a validated simulation framework, and lays the foundation for future deployment on physical platforms. This work provides a proof-of-concept for using deep reinforcement learning to enable autonomous manipulation and movement in simulated space-like conditions.

PPO ; Space ; manipulator ; reinforcement learning ; SKYWALKER ; ESA ; Isaac Lab ; Isaac Sim ; MIRROR ; Proof-of-concept ; microgravity

Download
View record in AAU Student Projects