Satellite forecasts and quantum-link studies

These figures were generated by executing the corrected PF Simulation r34 software. They show a historical replay of archived public CelesTrak orbital elements, not a live browser feed, spacecraft telemetry or a measured quantum experiment.

Orbit geometry

The orbit figure propagates the actual captured elements for ISS (NORAD 25544), TECHNOSAT (42829), and HST (20580) from 2026-09-16 18:32 to 20:32 UTC, with 121 samples per object. SGP4/WGS72 generates TEME position/velocity states; r34's Astropy conversion produces GCRS states. The Earth sphere is drawn at a radius of 6,378.137 km. The rendered curves join saved computed states; no trajectories were fabricated.

The three spacecraft supply example orbital geometry. Their assigned emitter/detector roles do not imply that the real spacecraft carry the modeled quantum equipment.

Matched orbit and photon/quantum comparison

The residual plot is a separate 120-second replay starting at the same UTC epoch. PF initializes a DOP853/J2 numerical orbit at each object's exact SGP4 GCRS position and velocity. Thirteen orbital epochs per object show the norm of numerical-minus-SGP4 position differences. The largest differences are 2.666875 m (ISS), 3.063653 m (TECHNOSAT), and 3.201510 m (HST). These compare two predictions and do not measure either prediction's physical accuracy.

Both orbit branches feed the real flat-SR moving-endpoint photon and quantum-channel workflow. Each branch executes three scenarios at three epochs: 18 scenario epochs, all with resolved propagation and Earth-clear geometry. Both numerical-validation reports passed. Scenario 3 uses Scenario 2's exact physical events and applies the pairwise PF analysis; no changed experiment is implied by changing the frame analysis.

The declared channel retains source depolarization 0.005, with phase noise and storage disabled. Conditional fidelity is 0.996250 and purity 0.99251875 throughout this example. Finite-count sampling is disabled. These values are conditional model predictions; they are not measured quantum-satellite performance or evidence of a PF advantage. The example does not execute a QKD exchange or a calibrated free-space optical link budget.

Reproduce the comparison

Install r34 and its tracking dependencies. Save config_satellite_gallery_replay.json in the extracted r34 package root, alongside run_comparison.py, so its paths resolve to the supplied historical fixtures and timing tables. From that directory run:

python run_comparison.py --config config_satellite_gallery_replay.json --output satellite_gallery_replay

Use a new output directory. The configuration is r34's canonical replay preset with only the comparison horizon changed to 120 s and orbit sample count to 13. The full two-hour orbit figure is reproduced by the included render_satellite_gallery.py script using the same captured elements; this plotting script also reads the saved 120-second comparison for the residual figure.

Exact public GP payloads, retrieval times, element epochs, source URLs, timing tables, and hashes remain in the archived run's inputs/ and session_manifest.json. At the historical replay epoch all records satisfy the configured 7-day age cutoff, without stale-data or degraded-IERS overrides.

Current-UTC capability

The installed r34 application also supports live current-UTC updates with network access. It refreshes CelesTrak GP inputs on a configurable cadence (2 hours by default) and propagates states at each requested epoch. This capability is implemented and covered by provider/policy tests; a new live network acquisition was not performed for these gallery figures.

For research planning, the value is controlled comparison of orbit and quantum-channel assumptions with replayable evidence. The result is not an operational mission-accuracy certification, measured photon experiment, or performance guarantee.