The steady increase in the number of space debris objects in Earth orbit demands unprecedented
solutions to ensure safe and sustainable access to space. In this context, Launch Collision Avoidance
(LCOLA) analyses are gaining increasing importance, complementing the well-established framework of
Space Traffic Management in safeguarding orbital operations.
This work presents an evolutionary topology-based methodology for LCOLA and its integration into
ASSET (AScent SafETy), DLR-GSOC’s decision-support tool for screening launch trajectories against
resident space objects. The known current operational LCOLA implementations rely on dense, discre-
tised liftoff sampling; while robust, these approaches are computationally expensive and may miss risky
conjunctions between samples.
To overcome these limitations, LCOLA is reformulated as a two-dimensional problem that minimizes the
primary–secondary distance in the frame of liftoff epoch (t₀) and Mission Elapsed Time (MET). A discrete
analysis identifies multiple (t₀, MET) intervals, each containing one local minimum of the distance, while
an optimization-based Time of Closest Approach (TCA) search using Differential Evolution (DE) is
performed in each identified interval containing a potential risky conjunction. This approach overcomes
state-of-the-art difficulties in mapping MET at the TCA back to the corresponding liftoff epoch, as both
variables are optimized and computed simultaneously, thereby achieving a direct relation.
Given the local minima identified through the DE optimization, two topology-based launch-closure com-
putation algorithms are implemented to determine the liftoff epochs around a risky TCA that yield a
collision risk (in terms of probability of collision PoC, miss distance, or both) exceeding a user-defined
threshold. A compact summary visualization overlays blackout windows with per-conjunction PoC and
miss-distance traces for operator decision support.
Validation on a representative test case shows that the evolutionary topology LCOLA recovers 100% of
risky conjunctions identified by the legacy discretized ASSET, while systematically refining TCAs and
miss distances. CPU time is the greatest advantage of this method as it achieves a reduction of more
than 90% when compared to the discretized LCOLA required for comparable accuracy.
Furthermore, rigorous verification using the operational COLA software (DLR-GSOC’s collision avoidance
tool) demonstrated accuracy in (t₀, MET) greater than 1 millisecond for each conjunction and launch
window closure.
The methodology therefore provides continuous launch-window coverage, realistic PoC-based risk quan-
tification, and a strong operational speedup, thus enabling DLR-GSOC to provide more flexible, accurate,
and timely LCOLA support for safe access to space.