Carry the retardation denominators in the radar Doppler model - #538
Merged
Conversation
tests/data/radar_synthetic.json was produced by a script that lived outside the repository, so the fixture could not be regenerated by anyone else and quietly went stale whenever the radar model changed. Add it under tools/ as it stood. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The Doppler observable is c d(tau)/d(t_receive). Differentiating the two
implicit light-time equations leaves a denominator on each leg:
dt_bounce /dt_receive = (c + rho_down.v_receiver)/(c + rho_down.v_asteroid)
dt_transmit/dt_receive = dt_bounce/dt_receive
* (c - rho_up.v_asteroid)/(c - rho_up.v_transmitter)
range rate = c (1 - dt_transmit/dt_receive)
To first order that is the instantaneous sum of the two one-way range rates,
which is what the model computed. The omitted term is a fractional error of
order rho.v/c. That is negligible for most targets and not at all negligible
for a fast one: on (6489) Golevka's 1995 apparition the line-of-sight rate
reaches 6 km/s, making it 5 to 7 Hz against stated uncertainties of 0.09 to
0.40 Hz -- a coherent, one-signed bias on every Doppler row.
The same denominator was already being applied to the range partials, as
ltdenom, so the Jacobian and the residual disagreed.
Both fixture generators encoded the old model and are updated with it, and
tests/data/radar_synthetic.json is regenerated: only the Doppler values move,
by 6e-7 to 1.7e-5 fractionally, and the delays are untouched.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
It was written outside the repository and so never saw the formatter. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #537.
On (6489) Golevka, with all 40 radar observations: reduced chi-square 6.53 → 0.79, the delay rows' rms(residual/sigma) 3.39 → 0.54, and the two centre-of-mass Doppler rows go from 63σ to 0.02σ. Aten improves slightly; Nyx is unchanged — as expected for a term that scales with range rate.
Both fixture generators encoded the old model. One lived outside the repository, which is why it went stale; the first commit brings it in under
tools/unchanged, the second corrects it with everything else. Only the Doppler values inradar_synthetic.jsonmove, by 6e-7 to 1.7e-5 fractionally; the delays are untouched. Full suite green (549 passed).The A2 figures quoted in #536 need that PR too — the non-grav solve cannot move A2 without it.
🤖 Generated with Claude Code