Enhancement of an AI-driven space situational awareness platform for more robust predictions

As transportation to space becomes more accessible, space debris pose increasing risk to operational satellites. Objects orbiting the earth can have a detrimental effect on space travel and threaten the
spacecraft and its personnel. These objects can be anything from active and passive satellites, orbital debris, space junk, asteroids and fragments from their disintegration and collisions. The first step toward
mitigating the adverse effects of these objects is to be able to predict their motion. Since the motions of these objects are very complex and not independent of each other, conventional techniques and directly
derived physical models fail. In this project, machine learning techniques are used to predict the motions of objects in space. Because machine learning techniques rely on a learning process, their accuracy can
wildly vary. This project specifically focuses on improving the prediction accuracy of two techniques implemented by Columbiad Launch Services Inc., by enhancing the data, and the training procedure that
leads to the machine learning.

Faculty Supervisor:

Nasser Lashgarian Azad

Student:

Mehran Zamani Abnili

Partner:

Columbiad Launch Services

Discipline:

Engineering - mechanical

Sector:

Professional, scientific and technical services

University:

University of Waterloo