FROM SIMULATION TO TESTING: CHALLENGES IN VALIDATING VISION-BASED SYSTEMS IN ADR AND OOS SCENARIOS

The estimation of the relative position and attitude (pose) of an uncooperative object with respect to a servicer spacecraft is a critical navigation task during proximity operations such as fly-around, Inspections, and close-approaches in Active Debris Removal (ADR) and On-Orbit Servicing (OOS) missions.

This seminar explores the key components behind the design, verification and validation of vision-based relative navigation systems in Rendezvous and Proximity Operations scenarios typical of these ADR/IOS missions, with special focus given to the pose estimation task and its critical challenges when tested in adverse orbital conditions and scenarios.

While giving strong emphasis on the improved robustness of novel image processing and relative navigation architectures based on Convolutional Neural Networks (CNN), the limitations and drawbacks of the validation of current navigation schemes with synthetic images are further discussed, highlighting the need to extend a purely synthetic validation into more comprehensive on-ground validations relying on more representative space imagery generated in laboratory test facilities.

Speaker:

Dr. Lorenzo Pasqualetto Cassinis is GNC Lead Engineer at Clearspace. In 2022, he obtained a Ph.D. from the Space System Engineering department at TU Delft. During his Ph.D., he was a researcher in the GNC section of ESA ESTEC, jointly with the space robotics team of Airbus Defence and Space, and a Visiting Researcher at Stanford's Space Rendezvous Laboratory. From December 2017 until May 2018, he also worked at GMV on the validation of the GNC software for the PROBA-3 Mission.

4.10.2024