Settings

Pc action threshold

Above this, an event is critical and a maneuver solution is generated. 1×10⁻⁴ is the common operational value.

Pc screening floor

Below this, events are logged but stay off the console.

Miss distance floor

Screen every approach closer than this regardless of Pc, so a near miss with an implausibly small covariance still surfaces.

Pc formulation

Foster and Chan are 2-D analytic methods valid for short, rectilinear hypervelocity encounters. Alfano handles longer encounters; Monte Carlo is the fallback when relative motion is non-linear.

Hard-body radius

Combined uses the sum of both circumscribing radii, inferred from RCS where no dimensions are on file. A fixed value is the conservative fallback for uncharacterized secondaries.

Covariance scale factor

Multiplier applied to the delivered covariance. Values above 1.0 inflate an optimistic covariance; a scaled covariance changes Pc non-monotonically.

Dilution handling

Pc rises then falls as covariance shrinks, so a small Pc from a large covariance is not the same as a small Pc from a tight one. Flagging surfaces events sitting in the dilution region; max-Pc reports the peak the event could still reach.

Propagator

SP integrates numerically against the force model below and carries a real covariance. SGP4 is the analytic mean-element theory a TLE encodes — faster, but with no formal covariance of its own.

Geopotential degree and order

J2 alone captures nodal regression and apsidal rotation. Higher orders matter for resonance and for repeat-ground-track orbits.

Atmosphere model

Drag is the dominant error source below ~700 km and the reason along-track covariance grows fastest. JB2008 responds better to solar activity; Harris-Priester is the cheap static option.

Solar radiation pressure

Matters most for high area-to-mass objects — deployed panels, insulation fragments, and anything uncharacterized.

Third-body perturbations

Lunisolar terms are small in LEO over a 72-hour screening window but grow with altitude.

Radial half-width

The RIC box each propagated state is screened against. Radial is kept tight because radial position is the best-determined axis.

In-track half-width

Widest axis by design — along-track is where drag and timing error accumulate, so the volume is elongated along the velocity vector.

Cross-track half-width

Set by plane-determination accuracy; usually matched to the in-track width for simplicity.

Command lead time

Upload, propagation, and ground-station contact margin subtracted from TCA to set the decision deadline on every event.

Burn model

Impulsive is valid when the burn is short against the orbital period. Electric propulsion needs a finite-burn model — thrust spans a meaningful arc, so the achieved delta-v is direction-averaged.

Preferred burn axis

In-track is the cheapest way to buy miss distance: a period change accumulates along-track separation as 3·Δv·t, so lead time does the work. Radial and cross-track act immediately but cost far more delta-v for the same separation.

18 SDS · Space-Track

radar · range, azimuth, elevation · 1σ 250 m

LeoLabs Radar Network

radar · range, range-rate, azimuth, elevation · 1σ 65 m

COMSPOC ASTRIAGraph

commercial · fused range, angles, operator ephemeris · 1σ 90 m

Slingshot Optical

optical · right ascension, declination · 1σ 480 m

ESA SST

radar · range, azimuth, elevation · 1σ 310 m

Fleet GNSS Downlink

onboard · GNSS pseudorange, carrier phase · 1σ 8 m