Ryan Shepperd, Iridium
https://doi.org/10.64861/TBUN9062
Keywords: conjunction risk assessment, conjunction risk mitigation, conjunction metrics, space traffic coordination, space traffic capacity
Abstract:
Seemingly easy metrics of the number of Conjunction Data Messages (CDMs) generated, conjunction warnings issued, and collision mitigation maneuvers executed, dominate policy discussions. Unfortunately, these metrics are often used out of context and are not normalized to such changing factors as screening volumes, reaction thresholds, and actual risk. Policy makers may not have a clear picture if they miss the context. On a different level, operators execute collision mitigation maneuvers based on chosen thresholds but struggle to measure the summed efficacy of those maneuvers. The aggregate probability of all conjunctions not mitigated and the aggregate of the remaining probability of all mitigated conjunctions contains the well-known problem of probability dilution, the assumption of independence between conjunctions, and the Gaussian and other assumptions built into the chosen probability method.
At both the policy and operational levels, good methods are needed to answer questions of what the current space traffic capacity is and how well we are controlling it. Much progress has been made in evolutionary debris models and virtual models for space traffic analysis such as MIT’s MOCAT and Georgia Tech’s VESTA with varying levels of realism. This paper aims to add to and inform those efforts with the practical experience of the Iridium constellation.
Measurable influences on risk and the degree of mitigation include
The conjunction consequences. Some operators, such as Iridium, weigh conjunction consequences in the mitigation strategy using the NASA EVOLVE break-up model as implemented by NASA CARA, split between those secondary objects with sufficient kinetic energy to create a large debris cloud of the primary and those that do not. Not all conjunctions pose the same consequences, and not all screening sources contain the same proportion of catastrophic and non-catastrophic conjunctions.
The probability of detection. The Sweetser probability of detection is the likelihood of registering a probability of collision above an operator’s chosen threshold. The baseline distribution of this metric can differ between screening sources and as will be shown through Iridium’s experience, is the key concerning metric affected by space weather in LEO.
Screening timespan and measurement cadence. The probability of detection obviously falls with longer propagation times, and so different operators and screening providers consider a different look ahead, usually between 3 and 7 days in LEO. Similarly, measurement cadence adds to propagation time between measurement epoch and TCA. Both also affect proactive conjunction mitigation via routine station-keeping. These maneuvers seek to minimize the aggregate probability of collision regardless of emergency mitigation threshold, and so longer spans and quicker cadence can improve that aim.
Coupled versus decoupled interaction between maneuverable satellites. Both the degree and type of cooperation and coordination between maneuverable satellites influence risk. On the obvious sides of the spectrum, there are those who share no information versus those who share all ephemerides, maneuver plans, and have a maneuver deconfliction system in place. However, decoupled cooperation is another method for safety by design. By ensuring control boxes about guidance targets do not intersect, real-time maneuver adjudication isn’t necessary or can be unilateral. A transiting satellite, for example, knows it can safely avoid a station-keeping satellite by planning a trajectory around its control box with no further communication required. Coupled interactions, on the other hand, represent a necessary subtraction from the trade space of avoiding other debris and satellites. The Iridium experience shows that in dense environments, mitigation maneuvers must often contend with other debris in the trade space. Increased interactions with other maneuverable satellites within that trade space blocks options and necessitates communication to avoid debris or third-party objects.
Residual risk after mitigation. All these metrics lead to the quintessential question of a satellite collision mitigation engineer: how well are we doing? The aggregate probability of collision derived from the CDMs overestimates the risk. Without maneuvers, the aggregate collision probability is best derived from the object flux. Models such as NASA ORDEM can provide this flux information as can techniques like the more recently developed Outer Space Institute’s CRASH Clock, using TLE data from the public catalog. However, for operators already on orbit, the tracked debris flux is better measured from the encounter rate of CDMs within a projected area sufficiently larger than maneuver effects on the flux, and good estimates of the hard body area. With a no-maneuver baseline established from the encounter rate, the reduction in probability of collision knowledge from the CDM must then be reconciled with the encounter rate.
Evolutionary debris and space traffic coordination models become more useful to evaluate strategy in satellite operations if they include these components. If more grounded in practical operational experience, such models could also better inform policy makers. These metrics observed by the Iridium constellation, along with suggested means to model them, will be discussed.
Date of Conference: September 15-18, 2026
Track: Conjunction/RPO