Measuring trajectory error and checking a heliocentric gravity correction against NASA SPICE.
My contributions
I added position- and velocity-error telemetry to Elodin's Voyager example, then contributed an optional heliocentric relative gravity model. Those changes were merged in PRs #769 and #789.
The comparison
Both saved runs completed 144 hourly steps. In this case, final position error against SPICE fell from 28.179 km in Chapter 1 to 0.840 km in Chapter 2.
The comparison uses the physical completed-step times; the database timestamps were one hour early. These are results for one simulation case, not a claim of navigation-grade accuracy.
