Research/ SPACE AI & AUTONOMY
Study on Model-based/Data-driven Reduced-Order Model for Autonomous Spacecraft
Seiji Tsutsumi
Rendezvous, Proximity Ops & DockingOpenAlex keyword: spacecraft docking · Learning-based MethodsOpenAlex topic: Model Reduction and Neural Networks
Due to the increasing needs for front-loading design and autonomous operation of spacecraft and test facility, the reduced-order models based on model-based/data-driven approaches are drawing attention. This article presents four examples of the development and application of the reduced-order model, 1) satellite thermal design, 2) spacecraft docking system design, 3) generation of training data for the health management of a reusable rocket engine, 4) operation of the wind tunnel. The existing challenges for the reduced-order model found through the development and application are introduced.