Geodesic
Application of machine learning methods for the classification of asteroid resonance motion
Vestnik Tomskogo gosudarstvennogo universiteta. Matematika i mehanika, no. 76 (2022), pp. 87-100 Cet article a éte moissonné depuis la source Math-Net.Ru

Voir la notice de l'article

When studying the resonant asteroid dynamics, it is necessary to classify time series of critical arguments on circulation, libration, or mixed case depending on their behavior. It is logical to use modern methods of machine learning to automatize this process. Earlier, a similar problem was solved for artificial satellites of the Earth. The purpose of this paper is to adapt the software attended for distinguishing resonant and nonresonant motion of satellites to solving asteroid dynamics problems. To achieve this goal, it is necessary to modify the program code and to train the created model on time rows obtained during the study of the asteroid orbital evolution. Operation of the modified software can be divided into three stages. At the first stage, to simplify the model-classifier, we make coding of time series of asteroid resonant arguments by vectors of lower dimension using an artificial neural network - an autoencoder. The second stage includes automatic clustering time series of asteroid resonant arguments by the HDBSCAN method (Hierarchical Density-Based Spatial Clustering of Applications with Noise) and their manual labeling to learn the classifier. At the third stage, based on the obtained training set, the artificial neural network-classifier is learned. The results of the classifier operation are estimated by visual comparison of graphs of the time series and received assessments. We may conclude that the classifier works correctly in most cases; some inaccuracies are observed in case of extreme amplitude and in the mixed case when libration passes to circulation.
Keywords: asteroid, mean motion resonance, machine learning, artificial neural network. 
@article{VTGU_2022_76_a6,
     author = {T. Yu. Galushina and E. A. Nikolaeva and D. S. Krasavin and O. N. Lenter},
     title = {Application of machine learning methods for the classification of asteroid resonance motion},
     journal = {Vestnik Tomskogo gosudarstvennogo universiteta. Matematika i mehanika},
     pages = {87--100},
     year = {2022},
     number = {76},
     language = {ru},
     url = {http://geodesic.mathdoc.fr/item/VTGU_2022_76_a6/}
}
TY  - JOUR
AU  - T. Yu. Galushina
AU  - E. A. Nikolaeva
AU  - D. S. Krasavin
AU  - O. N. Lenter
TI  - Application of machine learning methods for the classification of asteroid resonance motion
JO  - Vestnik Tomskogo gosudarstvennogo universiteta. Matematika i mehanika
PY  - 2022
SP  - 87
EP  - 100
IS  - 76
UR  - http://geodesic.mathdoc.fr/item/VTGU_2022_76_a6/
LA  - ru
ID  - VTGU_2022_76_a6
ER  - 
%0 Journal Article
%A T. Yu. Galushina
%A E. A. Nikolaeva
%A D. S. Krasavin
%A O. N. Lenter
%T Application of machine learning methods for the classification of asteroid resonance motion
%J Vestnik Tomskogo gosudarstvennogo universiteta. Matematika i mehanika
%D 2022
%P 87-100
%N 76
%U http://geodesic.mathdoc.fr/item/VTGU_2022_76_a6/
%G ru
%F VTGU_2022_76_a6
T. Yu. Galushina; E. A. Nikolaeva; D. S. Krasavin; O. N. Lenter. Application of machine learning methods for the classification of asteroid resonance motion. Vestnik Tomskogo gosudarstvennogo universiteta. Matematika i mehanika, no. 76 (2022), pp. 87-100. http://geodesic.mathdoc.fr/item/VTGU_2022_76_a6/

[1] B.M. Shustov, L.V. Rykhlova (red.), Asteroidno-kometnaya opasnost: vchera, segodnya, zavtra, Fizmatlit, M., 2010, 384 pp.

[2] Galushina T.Yu., “Orbitalnye i vekovye rezonansy v dvizhenii asteroidov, sblizhayuschikhsya s Zemlei”, Fizika kosmosa, tr. 49-i Mezhdunar. studencheskoi nauch. konf. (Ekaterinburg, 27-31 yanvarya 2020), UrFU, Ekaterinburg, 2020, 6–18

[3] Krasavin D.S., Aleksandrova A.G., Tomilova I.V., “Primenenie iskusstvennykh neironnykh setei v zadachakh analiza dinamicheskoi struktury oblastei okolozemnogo orbitalnogo prostranstva”, Izvestiya vuzov. Fizika, 63:3 (2020), 70–75

[4] Krasavin D.S., Aleksandrova A.G., Tomilova I.V., “Primenenie iskusstvennykh neironnykh setei v issledovanii dinamicheskoi struktury okolozemnogo orbitalnogo prostranstva”, Izvestiya vuzov. Fizika, 64:10 (2021), 38–43

[5] Grebenikov E.A., Ryabov Yu.A., Rezonansy i malye znamenateli v nebesnoi mekhanike, Nauka, M., 1978, 128 pp. | MR

[6] Geron A., Hands-On Machine Learning with Scikit-Leam, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems, 2nd ed., O'Reilly Media, Sebastopol, 2019, 484 pp.

[7] Burkov A., The Hundred-Page Machine Learning Book, 2019, 152 pp. (accessed: 27.07.2021) http://ema.cri-info.cm/wp-content/uploads/2019/07/2019BurkovTheHundred-pageMachineLearning.pdf

[8] Nilsson N.J., Introduction to machine learning (accessed: 26.04.2021) http://robotics.stanford.edu/people/nilsson/MLBOOK.pdf

[9] Parameter Selection for HDBSCAN : HDBSCAN Clustering Library Documentation (accessed: 09.01.2021) https://hdbscan.readthedocs.io/en/latest/parameter_selection.html