Kerstyn Auman, The Aerospace Corporation; Timothy S. Murphy, The Aerospace Corporation; Eric George, The Aerospace Corporation; Patrick Kelly, The Aerospace Corporation; Carson Coursey, The Aerospace Corporation; David Goldstein, SpaceX; Ian Ruh, SpaceX; Kate Hymes, SpaceX
Keywords: Space Situational Awareness,SSA,Conjunction Assessment,Collision Avoidance,CA,conjunction screening,validation,algorithm,Office of Space Commerce,OSC,Traffic Coordination System for Space,TraCSS,SpaceX
Abstract:
Validation Methodology for TraCSS Conjunction Assessment
Kerstyn Auman, Timothy S. Murphy
The Aerospace Corporation
Sandra Magnus
Office of Space Commerce
David Goldstein, Ian Ruh, Kate Hymes
SpaceX
Space Policy Directive 3 calls for the creation and maintenance of a civil commercial space traffic coordination (STC) system by the Department of Commerce (DOC). The operational solution, the Traffic Coordination System for Space (TraCSS), is currently in beta testing and has a variety of live and upcoming capabilities. One key capability of this system is conjunction assessment (CA), that is, the process of identifying if and when two space objects will pass closely by each other. Such close approaches result in a conjunction which is then typically subject to a more intensive risk assessment. The space safety mission requires a reliable, robust, efficient, and transparent CA process which can handle the demands of an ever-increasing space object population. The process of finding and evaluating an algorithm which fits this description is non-trivial, including requirements for algorithmic accuracy, computational efficiency, integrability, and acquisitional and operational cost.
This paper focuses on the specific technical challenge of ensuring algorithmic accuracy by means of a validation and verification, specifically on the conjunction finding portion of the CA screening process. Limiting the scope to exclude a validation of the probability of collision calculation is due to the likelihood that TraCSS will implement the NASA Conjunction Assessment Risk Analysis (CARA) department’s PcMultiStep algorithm, which can handle all conjunction encounter regions. As a research activity to help design the TraCSS system, the Aerospace Corporation worked directly with SpaceX’s Starlink team to validate their autonomous CA algorithm. The result of this process was a positive validation of the SpaceX methodology in low Earth Orbit (LEO) along with a variety of lessons learned during the iterations of the testing. This process helped both shape a more general process for assessing CA accuracy and identify common pitfalls and sources of error in CA, including interpolation and root-finding. This paper will present these results in the context of a set of requirements, both algorithmic and systematic, for a CA tool.
A dataset which can be used to evaluate an SSA provider’s CA algorithm accuracy is equally as important. This dataset must consist of a comprehensive collection of existing, potential, and physically possible orbit geometries to fully cover the trade-space of the current and future space environment and be formatted to mirror the standardized input expected by TraCSS. The dataset must be comprised of Consultative Committee for Space Data Systems (CCSDS) Orbit Comprehensive Message (OCM) ephemerides of different lengths relating to the following cases: maneuvering objects, heliocentric trajectories, very low relative velocity conjunctions such as those between two adjacent constellation objects, very high relative velocity conjunctions such as those between a prograde and retrograde object, numerically anomalous orbits such as those in equatorial orbits, and more. The TraCSS team will release a public dataset alongside this paper derived from the CA tool requirements developed during the SpaceX validation.
This paper will be composed of the following sections:
The conjunction screening validation methodology used on the SpaceX CA tool.
Description of the expected CA Tool requirements.
The process for generating a dataset, using a recent catalog snapshot as a starting point and supplementing it with ephemerides generated both from historical conjunctions of interest and fictitious objects when necessary.
A guide for using the results of this dataset to validate a specific CA tool.
BASIC RESULTS
The SpaceX algorithm was validated using a snapshot from the 18th Space Defense Squadron’s high accuracy catalog (HAC), which consists primarily of low earth orbit (LEO) objects, but does include a small variety of highly elliptical orbit (HEO) and geosynchronous (GEO) objects. While the SpaceX algorithm ultimately detects all the same conjunctions as the trusted internal provider within the accepted level of accuracy, the limited number of HEOs and GEOs in the test set precluded the validation from extending to all regimes. However, more rigorous testing will be required to approve an algorithm for use in the TraCSS system, which will conduct conjunction assessment for all trackable Earth-orbiting satellites.
The accuracy requirements for a conjunction screening algorithm are fairly straightforward, with the most important one being finding all the same events as a trusted provider to a desired level of accuracy. For very low relative velocity events (
Event Type
Specification
TCA Requirement (s)
Very Low Relative Velocity (VLRV)
= 50 m/s
0.001
The differences in TCA between the conjunctions found by both SpaceX and the internal provider are compared in relation to the relative velocity between two conjuncting objects. For conjunctions which have a relative velocity greater than 50 m/s, even the largest TCA differences are on the order of milliseconds for this category. Alternatively, the TCA differences between 10 and 50 m/s relative velocity are on the order of centi-seconds, with the only super-second differences occurring at less than 10 m/s relative velocity.
Date of Conference: September 16-19, 2025
Track: Conjunction/RPO