Cosmic ResearchCosmic ResearchКосмические исследования0023-42063034-5502The Russian Academy of Sciences69231510.31857/S0023420625040079qnnfljArticlesСтатьиResearch ArticleA Recursive Partitioning for Anomaly Detection in Tracking Satellite DataМетод рекурсивного разбиения для обнаружения аномалий в траекторных спутниковых данныхZapevalinP. R.ЗапевалинП. Р.pav9981@yandex.ruAstro Space Center of P.N. Lebedev Physics InstituteАстрокосмический центр Физического института имени П.Н. Лебедева РАН15082025634VOL 63, NO4 (2025)ТОМ 63, №4 (2025)42343707102025Copyright ©; 2025, Russian Academy of SciencesCopyright ©; 2025, Российская академия наук2025Russian Academy of SciencesРоссийская академия наукhttps://journals.eco-vector.com/0023-4206/article/view/692315

This study presents a method for detecting anomalous measurements in the trajectory data of spacecraft, based on recursive partitioning of the time series of observations. This method analyzes the standard deviation of the data, effectively identifying anomalous measurements characterized by elevated noise levels. A significant advantage of this approach is the lack of requirement for prior knowledge of the initial orbital approximation and the absence of a need for pre-training. It has been tested on synthetic datasets with artificially introduced anomalies, as well as on real data from the “Spektr-R” spacecraft. The results demonstrated an accuracy of 96 % compared to other traditional anomaly detection methods. The algorithm of this method is applicable to various types of orbits and scales of observations. Its code is available for public use.

В настоящей работе представлен метод обнаружения аномальных измерений в траекторных данных космических аппаратов, основанный на рекурсивном разбиении временной последовательности наблюдений. Данный метод анализирует среднеквадратическое отклонение данных, эффективно выявляя аномальные измерения, характеризующиеся повышенным шумом. Его преимущество заключается в отсутствии необходимости знания начального приближения орбиты и предварительного обучения. Метод был протестирован на модельных данных с искусственно введенными аномалиями и на реальных данных космического аппарата “Спектр-Р”. В сравнении с другими традиционными методами обнаружения аномалий на непосредственных наблюдениях данный подход продемонстрировал наименьший процент ложно-отбракованных измерений. Алгоритм этого метода подходит для различных типов орбит и масштабов наблюдений. Код алгоритма доступен для свободного использования.

Zapevalin P.Multi-GNSS Observations Simulation // Cosmic Research. 2024. V. 62. Iss. 5. P. 424–435.Zakhvatkin M.V., Andrianov A.S., Avdeev V.Yu. et al.RadioAstron orbit determination and evaluation of its results using correlation of space-VLBI observations //Advances in Space Research. 2020. V. 65(2). P. 798–812.Montenbruck O., Gill E.Satellite Tracking and Observation Models // Satellite Orbits: Models, Methods and Applications. Springer-Verlag Berlin Heidelberg, 2000. P. 193–232.Tuchin D.A.Selection of anomalous measurements during primary processing in orbit determination from the trajectory information of the ground station Cobalt-R // Keldysh Institute of Applied Mathematics Preprints. 2022. Iss. 58.Zakhvatkin M.V., Andrianov A.S., Avdeev V.Yu. et al.RadioAstron orbit determination and evaluation of its results using correlation of space-VLBI observations // Advances in Space Research. 2020. V. 65. Iss. 2. P. 798–812.Kovalenko I.D., Eismont N.A.Orbit design for the Spectrum-Roentgen-Gamma mission // Acta Astronautica. 2019. V. 160. P. 56–61.Эскобал П.Р.Методы определения орбит. М.: Мир, 1970. 471 с. (Перевод с англ. В.И. Ноздрина и В.М. Рудакова. Под ред. В.Г. Демина).Губанов В.С.Оценивание стохастических параметров обобщенным методом наименьших квадратов. СПб.: Институт прикладной астрономии РАН, 1994. 12 с. (Сообщения ИПА РАН; № 60).Tapley B.D., Schutz B.E., Born, G.H.Statistical Orbit Determination. Elsevier Academic Press, 2004. 563 p.Эльясберг П.Е. Определение движения по результатам измерений. М.: Наука, 1976. 416 с.Kalman R.A new approach to linear filtering and prediction problems transaction of the asme journal of basic // J. Basic Engineering (American Society of Mechanical Engineers). 1960. V. 82. Iss. 1. P. 35–45.Sorenson H.Kalman filtering : theory and application. IEEE Press, 1985. 457 p.D’Souza C.Fundamentals of Kalman Filtering and Estimation in Aerospace Engineering. NASA // Johnson Space Center. Houston, Texas, 2013.83 p.Coronel D., Guevara C.Anomaly Detection in the Uniaxial Vibration in Wind Turbines Applying Linear Regression and Z-score // Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops. Seattle WA, USA. 2024. P. 169–176.Chen D. and Meng D., Wang F. et al.A study of ionospheric anomaly detection before the August 14, 2021 Mw7.2 earthquake in Haiti based on sliding interquartile range method // Acta Geodaetica et Geophysica. 2023. V. 58. P. 539–551.Romo-Chavero M.A., Cantoral-Ceballos J.A., Pérez J.A. et al. Median Absolute Deviation for BGP Anomaly Detection // Future Internet. 2024. V. 16(5). Art.ID. 146.Grubbs F.E.Sample Criteria for Testing Outlying Observations // The Annals of Mathematical Statistics. 1950. 21(1). P. 27–58.Duchnowski R., Wiśniewski Z.Accuracy of the Hodges–Lehmann estimates computed by applying Monte Carlo simulations // Acta Geodaetica et Geophysica. 2016. V. 52. P. 511–525.Baireddy S., Desai S., Mathieson J. et al.Spacecraft Time-Series Anomaly Detection Using Transfer Learning // Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops. Nashville, TN, USA. 2021.Zapevalin P.R., Novoselov A., Zharov V.E. Artificial neural network for star tracker centroid computation // Advances in Space Research. 2022. V. 71. Iss. 9. P. 3917–3925.Herrmann L., Bieber M., Verhagen W. et al. Unmasking overestimation: a re-evaluation of deep anomaly detection in spacecraft telemetry // CEAS Space J. 2024. V. 16. P. 225–237.Cuéllar Carrillo S., Santos Penas M., Alonso F. et al. Explainable anomaly detection in spacecraft telemetry // Engineering Applications of Artificial Intelligence. V. 133(4). P. 1–15.Biswas G., Khorasgani H., Stanje G. et al. An Application of Data Driven Anomaly Identification to Spacecraft Telemetry Data // Annual Conference of the PHM Society. 2016. V. 8(1).Liu L., Tian L., Kang Zh. et al.Spacecraft Anomaly Detection with Attention Temporal Convolution Network // Neural Computing and Applications. 2023. V. 35. P. 9753–9761.Ré N.P., Popplewell M., Caudill M. et al. Transformers for Orbit Determination Anomaly Detection and Classification // Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops. Seattle, WA, USA. 2024. P. 6819–6827.Hundman K., Constantinou V., Laporte C. et al. Detecting Spacecraft Anomalies Using LSTMs and Nonparametric Dynamic Thresholding // Proc. 2018 ACM SIGSAC Conference on Computer and Communications Security. London, United Kingdom. 2018. P. 387–395.Montenbruck O., Gill E.Satellite Orbits. Models, Methods, and Applications. Springer-Verlag Berlin Heidelberg, 2000. 382 p.Kara I., Bazyey O.Using of Everhart’s method of 15, 17, 19 and 21st-order for computation of celestial bodies’ trajectories in the circumplanetary space // Astronomical School’s Report. 2009. V. 6(1–2). P. 155–157.Jin S., Cardellach E., Xie F. Introduction to GNSS // GNSS Remote Sensing. Remote Sensing and Digital Image Processing. V. 19. Springer, 2014. P. 3–16. ISBN: 978-94-007-7481-0. DOI: 10.1007/978-94-007-7482-7.Gulyaeva T. Investigation of total electron content deduced from TOPEX-JASON and GPS-IONEX maps over land and oceans // Proc. 37th COSPAR Scientific Assembly. Montréal, Canada. 2008. V. 37. P. 31–36.Mendes V., Langley R.Tropospheric zenith delay prediction accuracy for airborne GPS high-precision positioning // Proc. Institute of Navigation 54th Annual Meeting. 1998. P. 337–347.Boehm J., Niell A., Tregoning P. et al.Global Mapping Function (GMF): A new empirical mapping function based on numerical weather model data // Geophysical Research Letters. 2006. V. 33(7).Kutschera M., Zajiczek W.Shapiro effect for relativistic particles — testing General Relativity in a new window // Acta Physica Polonica Series B. 2009. V. 41(6).Zakhvatkin M., Ponomarev Yu., Stepanyants V.Navigation support for the RadioAstron mission // Cosmic Research. 2014. V. 52(4). P. 342–352.Тучин Д.А. Определение орбиты на борту космического аппарата // Известия РАН. Теория и системы управления. 2020. № 3. С. 126–147.Lam Q., Junker D., Anhalt D. et al. Analysis of an Extended Kalman Filter Based Orbit Determination System // Proc. AIAA Guidance, Navigation, and Control Conference. Toronto, Ontario, Canada. 2010.Luo J., Ying K., Bai L.Savitzky–Golay smoothing and differentiation filter for even number data // Signal Processing. 2005. V. 85(7). P. 1429–1434.Virtanen P., Gommers R., Oliphant T.E. et al.SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python // Nature Methods. 2020. V. 17. (Suppl. 1). P. 1–12.Dach R., Andritsch F., Arnold D. et al. Bernese GNSS Software Version 5.2. Bern: Astronomical Institute, University of Bern, 2015.Rudnitskiy A., Mzhelskiy P.V., Shchurov M. et al. Analysis of orbital configurations for Millimetron space observatory // Acta Astronautica. 2022. V. 196(1). P. 29–41.