Landing a spacecraft on an extraterrestrial body is one of the most significant challenges. Landing a spacecraft other than the earth’s surface just by taking video footage from a camera is almost impossible. In recent years, AI algorithms have significantly improved, but to train deep learning models, massive amounts of datasets are required. On Kardashev scale we are not even a Type-I civilization which could harness all the energy that is available from a neighboring star, gathering and storing it to meet the energy demands of a growing population therefore collection of data while actual spacecraft landing on planet other than earth is not feasible at current state. But all this work like collection of training data is possible inside simulator. As simulator does not required actual fuel thus any amount of training data could be generated while landing a spacecraft in the simulator. Therefore in this research a simulation is created as realistic as possible considering required physics parameters. Data has been collected while landing a spacecraft, trained deep neural network, and deployed those DNN into simulator.
Deep Learning Approach for Autonomous Spacecraft Landing
Lect. Notes in Networks, Syst.
International Conference on Intelligent Systems Design and Applications ; 2023 ; Olten, Switzerland December 11, 2023 - December 13, 2023
25.07.2024
11 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Optical guidance for autonomous landing of spacecraft
IEEE | 1999
|Optical guidance for autonomous landing of spacecraft
Tema Archiv | 1999
|PAPERS - Optical Guidance for Autonomous Landing of Spacecraft
Online Contents | 1999
|