CAPSTONE Telemetry Shows Eclipse Timing Cuts Cislunar Orbit Uncertainty Without Ground Links

The arXiv preprint formulates eclipse entry and exit times as implicit zero-velocity measurements on second-order shadow cones derived from planetary ephemerides. These events are weighted by realistic timing uncertainty arising from telemetry sampling rate, Earth atmosphere refraction, and lunar topography, then ingested by an unscented Kalman filter. Both L1 halo simulations and January 2026 CAPSTONE flight data demonstrated measurable covariance reduction, particularly when ground-tracking updates were withheld.

Standard spherical-body shadow models omit atmospheric and topographic effects that shift contact times by up to several seconds; the authors propagate those physical residuals into the measurement covariance, preventing over. This approach requires only existing solar-array current telemetry and ephemeris data, eliminating new hardware.

Related work on X-ray pulsar navigation and optical horizon sensing shows similar opportunistic-measurement gains but demands specialized detectors. Eclipse timing therefore occupies a low-cost niche for near-term cislunar fleets where mass and power budgets preclude additional sensors.

Operational adoption hinges on demonstrating robustness across multiple eclipse seasons and integrating the observables into existing flight-software Kalman architectures.