Orbit Accuracy Experiments

Space Data Network · Evidence

Orbit accuracy, measured in the open.

Experiments on how accurate orbit propagation and catalog covariance are, against precise orbits. Every number here is computed by Space Data Network WebAssembly modules, and you can re-run them on this page, in your browser.

Live

Propagate A Real Orbit

Pick a satellite with a precise orbit and a force model. Your browser loads propagator/hpop, seeds it with the precise state, propagates three days and compares every sample with the precise orbit.

Loading the modules…

3D position error against the precise orbitRun a propagation to plot it.

V1 · HPOP Against Real Orbits

Each Force, Measured

HPOP seeded from 16 SLR states (LAGEOS-1/2, ETALON-1/2) and 96 GPS states, propagated three days and compared with the same precise orbits, force model by force model.

SLR spheresmedian 3D error, m
GPSmedian 3D error, m

Criteria

    Configurations

    Verification · HPOP Against Orekit 13.1

    Locked Down Force By Force

    72 cases on identical states, forces and constants: five orbits, ten force sets, and the IERS 2010, VCM and atmosphere additions. Each point is the largest position difference over a day of hourly samples. The reference is Orekit 13.1 with its de Sitter term evaluated in consistent frames. Pick a case and re-run it here. Four cases read space weather or JB2008 indices that are not republished here; they are listed as recorded only.

    Largest difference from Orekit over 24 hm, one point per case; the hairline is the case's tolerance

    VCM Parity · analysis/vcm-adapter

    A Vector Covariance Message, Read and Propagated

    An SP Vector Covariance Message read into a PRW request and propagated with its covariance. The format states no units for its covariance. Four SP messages show that its mean-motion row is relative (dn/n); with that reading the sigmas the adapter recomputes match the printed ones within 1 %. The ballistic coefficient’s row is read the same way, with the as-printed reading run beside it.

    The message

    
                
                

    What the adapter read

      Position sigmas along the propagationm, from HPOP’s 7×7 covariance with the ballistic coefficient B as a parameter; in-track under both readings of B’s row

      Experiments

      Plans First, Then Numbers

      Each experiment states its hypotheses, data windows and decision rules before its test data is read. Failures are reported with the same care as passes.

      Data And Provenance

      Every File, With Its Hash

      The module binaries this page runs, the precise orbits, Earth orientation and ephemeris it reads, and the results it shows. Element sets from Space-Track never leave the machine they were fetched to and are not here.