Ryan Shepperd, Iridium
Keywords: Conjunction assessment, collision avoidance, CA, COLA, conjunction metrics
Abstract:
First launched in 1997 with a vehicle refresh in 2017, the Iridium constellation has endured three solar cycles with the first peaking in 2003. The modern era of actionable collision assessment data and active collision avoidance only began in 2009, though, during solar cycle 24, whose activity was mild compared to historical cycles. The advent of a stronger cycle 25 renewed interest at Iridium in this key influence on space object tracking and, by extension, quantifying performance.
As a metric of space debris progress or regression, the number of conjunction data messages issued, and the number of collision avoidance maneuvers performed are often cited. The total number of conjunctions issued to an operator is dependent on the screening volume and predictive time span. The change in the number of conjunctions within this volume can be a direct measure of the local debris environment seen by a constellation if the volume is constant. However, the volume has not been constant, a caveat that can be missed when these numbers are reported. The volume has, in fact, grown over time per the request of operators who sought that enlargement both to decrease surprise conjunctions and to facilitate station keeping maneuver pre-screening. Today’s volumes thus include far more conjunctions than just those that could be a concern and should not factor in any assessment of false alarm rates.
Moreover, the reaction to those conjunctions contains additional nuances. Collision avoidance maneuvers are difficult to compare between operators, orbital regimes, and over time. At the first obvious level, perhaps the distinction sought in a comparison, different operators utilize not only different probability of collision thresholds but different probability of collision methods. Though the industry has converged on a best practice of 1e-4 as a minimum threshold, some operators use a more conservative threshold of 1e-5, and the action time threshold can be dependent on mission objectives and constraints. Further complicating a comparison, not every probability calculation is the same; methods with differing levels of conservatism and simplifications are used.
A third complexity in avoidance maneuver numbers is an approximate inverse relationship with solar activity. When solar activity is lower, the number of avoidance maneuvers has been higher; when solar activity is higher, the number of avoidance maneuvers is lower. This relationship has two contributions. First, the opportunistic minimization of collision probability through station keeping is greater during higher solar activity with higher atmospheric density. There are more station-keeping maneuvers and more opportunities. In a lower drag environment, the period between maneuvers can be long and any concerning conjunctions in those periods must be handled with a dedicated avoidance maneuver. Second, an inconvenient truth is just because an object is reliably tracked does not mean it can be avoided. Larger covariances do not produce collision probabilities that surpass oft-chosen thresholds. The largest covariances can also surpass a typical control box size, making the object’s position error impossible to avoid within mission constraints. Covariance size has a direct relationship to solar activity, or solar variability. The number of avoidance maneuvers is thus reduced during solar max due to the positive influence of opportunistic station-keeping and the negative influence of increased orbit propagation errors.
Without an effort to address the diversity of variables affecting these metrics, using them in comparisons, be it historical trending or between operators, is unsound. Ideally, collision avoidance performance should be evaluated among all maneuvers in comparison to the background risk and actionability, and the converse of the latter should be the emphasis of future improvements. In this assessment of Iridium’s count of avoidance maneuvers, the influences, limitations, and suggestions to achieve comparable metrics will be discussed.
Date of Conference: September 17-20, 2024
Track: Conjunction/RPO