Skip to main content

Risk Metrics

Conjunction events are evaluated using multiple metrics to quantify collision risk under uncertainty. Several Pc methods are available because each makes assumptions about the conjunction geometry, and Satcat selects the method whose limitations that geometry respects.


Time of Closest Approach (TCA)

The instant when two RSOs reach minimum separation, reported to millisecond precision.


Miss Distance (MD)

The relative separation between two RSOs at TCA, expressed as a vector in the RIC or VNB frame and derived from nominal state estimates.


Measurement Covariance

The position and velocity uncertainty of a state estimate, produced by orbit determination and expressed as a 6x6 matrix at the state's epoch. Covariance grows as a state is propagated forward, so a CDM issued days before TCA carries more uncertainty than one issued hours before.

Pc methods integrate the combined uncertainty over the hard body volume. Where covariance is missing for one object, Max Pc methods are used instead.


Mahalanobis Distance

Calculates how far the nominal miss distance lies from the center of the uncertainty distribution:

d_M = sqrt(Δrᵀ · Σ⁻¹ · Δr)

Where:
- Δr is the miss distance vector
- Σ is the combined 3x3 covariance matrix

Low values indicate that the nominal miss distance lies within the combined uncertainty, where collision risk is most significant.


Tau Bounds

The Tau bounds metric represents the time interval around TCA during which the positional uncertainty volumes of two RSOs overlap the combined hard-body radius. This interval is derived by projecting relative motion and covariance into the encounter plane and applying a statistical confidence threshold to identify when collision risk is non-negligible.


Probability of Collision

Probability of collision (Pc) estimates the likelihood that two objects will collide by integrating the combined uncertainty over the collision volume defined by their hard-body volumes. Different methods are used depending on the event geometry and the available data.

Pc Methods

MethodDescription
Foster-1992 (2D)Standard encounter plane assessment. Assumes Gaussian uncertainties, rectilinear motion, and no offset violations.
Hall (3D Nc)Used when 2D assumptions break down. Required for precise Pc in high-curvature or low-relative velocity cases.
SDMCSimplified-dynamics Monte Carlo using equinoctial sampling at TCA and two-body Keplerian dynamics
Frisbee (Max Pc)Conservative estimate used when only one object's covariance is available, or the Pc is in the dilution region.
Frisbee Max Pc

The Frisbee method estimates an upper bound by assuming a worst-case relative geometry when full covariance data is not available.


How Satcat Computes Risk Metrics

When a close approach is detected, Satcat computes risk metrics in this sequence:

  1. Compute TCA and nominal MD
  2. Validate and remediate NPD covariance
  3. Compute Mahalanobis distance
  4. Identify best HBR and update CDM if needed
  5. Determine Pc method:
    • Use Foster-1992 when 2D geometry assumptions are satisfied
    • Escalate to 3D Nc [Hall] when the 2D assumptions are violated
    • Apply SDMC when the 3D Nc method fails to converge or reports an extended violation
    • Use Frisbee Max Pc estimate if only partial covariance is available
  6. Flag dilution and report the Max Pc where applicable
  7. Recalculate metrics as new data are ingested

Pc Dilution

In high-uncertainty conjunctions a nominal Pc can understate the actual risk. Satcat identifies these dilution region conjunctions and reports Maximum Pc alongside nominal Pc.

See Pc Dilution for how dilution is recognised and what it means for a mitigation decision.


Hard Body Radius (HBR)

Collision probability estimates across Satcat depend heavily on the HBR values used for the primary and secondary RSOs. Satcat draws HBR from the following sources, in precedence order, highest first:

  1. Satcat Operators: Manually specified values from RSO O/Os
  2. NASA CARA: Values provided by NASA's Conjunction Assessment Risk Analysis group
  3. Satcat DB: Values from Satcat's internal database
  4. From CDM: Uses the HBR specified in the associated Conjunction Data Message
  5. Fallback Space-Track OBJECT_TYPE logic (ex: 5 m for Payloads)

The selected HBR directly impacts computed collision probabilities and should be reviewed carefully, especially when validating risk thresholds for operational decision-making. HBR source information is provided in Event page details and selectable in Maneuver Designer projects.