Overcoming Dilution of Collision Probability in Satellite Conjunction Analysis via Confidence Distribution

Area of Science:

• Space Situational Awareness
• Astrodynamics
• Probabilistic Inference

Background:

• Dilution of collision probability is a known issue in satellite conjunction analysis.
• The false confidence problem is another limitation, potentially linked to probability dilution.

Purpose of the Study:

• To distinguish between probability dilution and false confidence.
• To introduce a confidence distribution (CD) with a point mass at zero for satellite conjunction analysis.

Main Methods:

• Investigated a confidence distribution (CD) with a point mass at zero.
• Analyzed the role of the point mass in capturing data uncertainty.

Main Results:

• The CD resolves probability dilution and ambiguity in Neymanian confidence intervals.
• The CD addresses false confidence in satellite conjunction analysis.
• Confidence derived from the CD directly interprets as a p-value for collision risk hypothesis testing.

Conclusions:

• A confidence distribution with a point mass at zero is essential for accurate satellite conjunction analysis.
• This approach improves probabilistic inference and collision risk assessment.
• The method provides a direct p-value interpretation for hypothesis testing.