Abstract :
[en] This thesis presents the development and implementation of a high-fidelity framework for emulating on-orbit operations using lightweight manipulators in a ground-based environment, specifically within the Zero-G Lab. Here, "high-fidelity" refers to accurately reproducing the dynamic characteristics of in-space scenarios, such as the motion of free-floating space systems, to enable realistic physics-based behaviour. In this context, emulation involves using physical mockups—scaled or full-size—whose motion is driven by simulated dynamics and executed by robotic hardware to replicate specific in-space scenarios on Earth. On-ground emulation provides a controlled, cost-effective alternative to replicating complex microgravity conditions, enabling pre-deployment testing of satellite behaviors, robotic servicing, capture maneuvers, and other in-orbit operations. The work focuses on overcoming major challenges in robot-assisted emulation, particularly related to kinematic control, time-delay, and dynamic scaling. Although modern industrial robots are capable of precise motion, limitations to emulation fidelity arise from near-singularities in inverse kinematics, discrete-time hardware delays, and mismatches between full-scale spacecraft dynamics and scaled-down laboratory setups. To address these, a unified emulation framework was developed integrating robust motion control, time-delay compensation, and dynamic scaling. Specifically, the framework employs Virtual Forward Dynamics Models (VFDM) to achieve stable motion near kinematic singularities, passivity-based control strategies to counteract time-delay-induced instabilities, and analytical scaling laws to ensure dynamic equivalence between lab-based and orbital conditions. The system is implemented using open-source tools such as the Robot Operating System (ROS) to promote reproducibility and scalability. Experimental validation demonstrates that the framework can accurately replicate various on-orbit scenarios using lightweight, position-controlled robots. However, limitations such as sensor drift, imperfect inertia compensation, and the inability to fully replicate microgravity remain. Future work directions are identified, including enhanced force sensing, improved dynamic parameter identification, and adaptive scaling strategies. These aim to further advance reliable, high-fidelity robotic emulation platforms for future space missions.
