Artificial Neural Network Design for Tours of Multiple Asteroids

Viavattene, G. and Ceriotti, M. (2020) Artificial Neural Network Design for Tours of Multiple Asteroids. In: 15th International Conference on Hybrid Artificial Intelligence Systems (HAIS 2020), Gijón, Spain, 11-13 Nov 2020, pp. 751-762. ISBN 9783030617042 (doi: 10.1007/978-3-030-61705-9_63)

[img] Text
226174.pdf - Accepted Version



Designing multiple near-Earth asteroid (NEA) rendezvous missions is a complex global optimization problem, which involves the solution of a large combinatorial part to select the sequences of asteroids to visit. Given that more than 22,000 NEAs are known to date, trillions of permutations between asteroids need to be considered. This work develops a method based on Artificial Neural Networks (ANNs) to quickly estimate the cost and duration of low-thrust transfers between asteroids. The capability of the network to map the relationship between the characteristics of the departure and arrival orbits and the transfer cost and duration is studied. To this end, the optimal network architecture and hyper-parameters are identified for this application. An analysis of the type of orbit parametrization used as network inputs for best performance is performed. The ANN is employed within a sequence-search algorithm based on a tree-search method, which identifies multiple rendezvous sequences and selects those with lowest time of flight and propellant mass needed. To compute the full trajectory and control history, the sequences are subsequently optimized using an optimal control solver based on a pseudospectral method. The performance of the proposed methodology is assessed by investigating NEA sequences of interest. Results show that ANN can estimate the cost or duration of optimal low-thrust transfers with high accuracy, resulting into a mean relative error of less than 4%.

Item Type:Conference Proceedings
Additional Information:Part of the Lecture Notes in Computer Science book series (LNCS, volume 12344).
Glasgow Author(s) Enlighten ID:Ceriotti, Dr Matteo and Viavattene, Giulia
Authors: Viavattene, G., and Ceriotti, M.
College/School:College of Science and Engineering > School of Engineering > Systems Power and Energy
Copyright Holders:Copyright © Springer Nature Switzerland AG 2020
First Published:First published in HAIS: International Conference on Hybrid Artificial Intelligence Systems
Publisher Policy:Reproduced in accordance with the publisher copyright policy
Related URLs:

University Staff: Request a correction | Enlighten Editors: Update this record