Collision Risk-Aware Launch Planning

Abstract

Rapid expansion of satellite and debris populations in Low Earth Orbit (LEO) intensifies the risk posed to and by launch activities. Regulatory authorities require mitigation of this risk through Collision on Launch Avoidance (COLA) practices by imposing thresholds on the accepted probability of collision (PoC). While Conjunction Assessment (CA) for operational satellites is deeply analysed in the literature, limited research addresses the launch phase, where state-uncertainty predictions are large and irreducible.

Developed in collaboration with Rocket Factory Augsburg (RFA), this research set out to investigate a COLA methodology that reliably assesses collision risk under large trajectory uncertainties. This was achieved by evaluating the validity of traditional CA assumptions under launch conditions for a representative polar orbit trajectory and estimating the operational reliability of various PoC metrics in terms of missed detections and false alarms.

Results indicate that while standard CA assumptions generally hold, single-Gaussian uncertainty overestimates launch dispersions, causing traditional risk metrics to be affected by probability dilution. Conversely, modelling launcher uncertainty with Gaussian Mixture Models (GMMs) improves operational reliability. Ultimately, this work provides a structured and reproducible validation framework to assess collision risks during launch operations.