Collision Avoidance and Formation Flight

Published:

Role: Research and thesis supervision ยท Chair of Spacecraft Systems, Technical University of Munich

Conjunction handling today is decided on the ground, one operator at a time. As orbital density grows that pattern scales badly: the number of screenings rises faster than the number of satellites, and each unilateral manoeuvre perturbs the conjunction geometry for everyone else.

This work asks what changes when the spacecraft involved decide cooperatively which of them manoeuvres, and how. Treating the manoeuvre as a shared optimisation lets the cost be distributed - the satellite with propellant to spare moves, rather than the one that happened to be flagged - and cuts the communication overhead that centralised coordination imposes. The same machinery applies to formation reconfiguration, where the trigger is a change in mission geometry instead of a conjunction warning.

Related publication: Orbital Manoeuvring Optimization Techniques for Collision Avoidance through Decentralized Algorithms, 75th International Astronautical Congress, 2024.

Supervised theses

Adaptive Reconfiguration Framework for Decentralized Optimization in Spacecraft Formation Flight and Collision Avoidance

Developed a dynamic reconfiguration framework in MATLAB and STK that scores manoeuvre strategies on propellant, energy, timing and satellite availability. It showed that decentralised reconfiguration improves resource use and scalability across formation sizes, with collision avoidance as the main application.

Orbital Manoeuvring Optimization Techniques for Collision Avoidance through Cooperative Algorithms

Proposed a decentralised framework in which satellites decide cooperatively when and how to manoeuvre to avoid collisions. It optimises propellant, energy and data exchange while cutting the communication overhead of centralised control. The approach was validated against conventional thrust-based manoeuvres.

Orbital Manoeuvring Optimization Techniques for Collision Avoidance through Decentralized Algorithms, IAC 2024