LEVERAGING MACHINE LEARNING FOR ENHANCED SATELLITE TRACKING TO BOLSTER SPACE DOMAIN AWARENESS

Our modern society is more dependent on its assets in space now more than ever. For
example, the Global Positioning System (GPS) many rely on for navigation uses data from a
24-satellite constellation. Additionally, our current infrastructure for gas pumps, cell phones,
ATMs, traffic lights, weather data, etc. all depend on satellite data from various constel-
lations. As a result, it is increasingly necessary to accurately track and predict the space
domain. In this thesis, after discussing how space object tracking and object position pre-
diction is currently being done, I propose a machine learning-based approach to improving
the space object position prediction over the standard SGP4 method, which is limited in
prediction accuracy time to about 24 hours. Using this approach, we are able to show that
meaningful improvements over the standard SGP4 model can be achieved using a machine
learning model built based on a type of recurrent neural network called a long short term
memory model (LSTM). I also provide distance predictions for 4 different space objects over
time frames of 15 and 30 days. Future work in this area is likely to include extending and
validating this approach on additional satellites to construct a more general model, testing a
wider range of models to determine limits on accuracy across a broad range of time horizons,
and proposing similar methods less dependent on antiquated data formats like the TLE.