Enhance the TLE catalog through sharing machine learning models

Hao Peng, Xiaoli Bai, Lesley A. Weitz, Scott C. Kordella

Research output: Contribution to journalConference articlepeer-review

5 Scopus citations


We have recently developed a methodology to predict trajectories of resident space objects (RSOs) with higher accuracy than the current methods. The new method leverages accurate state-of-the-art physics models and uses ML methods to provide other necessary details that are unresolved theoretically. As an important next step towards establishing an information sharing capability, this paper proposes and studies a monitoring system of the ML approach to enhance the TLE catalog. The training, testing, and evaluation of the ML models require learning from a regularly updated dataset. And the trained ML models need to be examined by a monitoring system before its distribution, due to the statistical nature of the ML methods. The design, implementation, and analyzing of the monitoring system is presented in this paper. Simulation results show that the prototype design can track the performance of the trained ML models and the re-training of the ML models could significantly improve the performance of the ML models.

Original languageEnglish (US)
Article numberIAC-19_A6_7_10_x50036
JournalProceedings of the International Astronautical Congress, IAC
StatePublished - 2019
Event70th International Astronautical Congress, IAC 2019 - Washington, United States
Duration: Oct 21 2019Oct 25 2019

All Science Journal Classification (ASJC) codes

  • Aerospace Engineering
  • Astronomy and Astrophysics
  • Space and Planetary Science


  • Gaussian Processes
  • Information Sharing
  • Machine Learning
  • Monitored Machine Learning Database
  • Orbit Predictions
  • TLE catalog


Dive into the research topics of 'Enhance the TLE catalog through sharing machine learning models'. Together they form a unique fingerprint.

Cite this