Design an AI tool to predict satellite activity, dangerousness and failures

A satellite service provider manufactures its platform with the same bus for the constellation to reduce the development costs. The same type of mission as well as activity are expected from these satellites. In addition, same events could occur due to similar hardware and design quirks. Malfunctions and system failures can therefore be recurrent. To be able to foresee future activities and potential failure is important for Space Traffic Management (STM).


The main goals are to identify patterns in mission failures as well as sorting the satellites with respect to their dangerousness. This includes:

  • Understand and find the history of events / failures for the different types of satellite
  • Define a risk factor and find the most «dangerous» satellites
  • Train an algorithm and predict future failure
  • Mine and characterize different types of events

Required skills

For the good realization of the project, it is recommended that the student have:

  • General understanding of the orbital dynamics and satellite system engineering
  • Programming skills in Java or python and knowledge of AI and MT tools is a plus
  • Ability to appraise and adapt the project to fulfill the needs for the Space Domain and its partners
  • A sense of responsibility for the quality of the work to be used for everyday applications

Place of work

The student will have the opportunity to work in a stimulating environment with other students in different locations (Lausanne, Bern, Zurich) and at different observation sites depending on the needs of the projects.



The student is expected to deliver at the end of the project the following elements:

  • An automatized evaluation of the satellite safety in the form of reports and/or dynamic database
  • The source code developed during the project
  • A complete documentation of the development and methodology, including a description of the environment, the installation procedure and configuration
  • The results of the analyses with limitations and recommendations for further work