Colloquium: Antonio Lopez Rivera (Space Flight)
31 October 2024 13:30 - Location: Lecture Room C, Faculty of Aerospace Engineering, Kluyverweg 1, Delft | Add to my calendar
Reinforcement Learning for Multi-rendezvous Mission Design
The design of multi-target rendezvous trajectories, which see a spacecraft approaching a sequence of objects in orbit as efficiently (by some metric) as possible, is a challenging problem of critica! importance for Active Debris Removal (ADR) and cis-Lunar logistics. This research aims to assess the applicability and effectiveness of Neural Combinatorial Optimization (NCO) methods for the design of multi-rendezvous missions. The research shows that NCO approaches can achieve optimality gaps lower than 30% in ADR scenarios. A framework is presented for the design of multi-rendezous missions with strict feasibility guarantees. This framework is used to design multi-rendezvous tor and analyze the operational envelope of UARX Space's OS. Vehicle due to launch in 2025.
Supervisor: Marc Naeije