Use the transmitting antenna for the radar up leg (bistatic) - #534
Merged
Conversation
An Observation carried one station, used for both legs, so a bistatic measurement had its up leg evaluated at the antenna that received rather than the one that transmitted. On (6489) Golevka that is about 125 sigma per Doppler point -- observed/model 1.039-1.044 on bistatic rows against 0.999934 on monostatic ones -- and 20 of its 40 rows are bistatic. Observation now carries the transmitting antenna's state at the transmit epoch, and the up leg uses it directly. Where it is absent the receive station is extrapolated as before, so monostatic behaviour is unchanged. The transmit epoch does not have to be solved for: the round-trip delay is the observable, so it is t_receive - tau. Golevka's bistatic rows are all Doppler-only and carry no delay of their own, so tau is interpolated from the object's delay rows, which run smoothly from 46.5 to 47.2 s across the 1995 apparition. Evaluating the station rather than extrapolating to it also removes the Taylor truncation, measured at 2.2e-3 m/s in velocity at that round trip. Measured on Golevka, all 40 radar observations: bistatic, receive station extrapolated flag 2, reduced chi-square 1.487e7 bistatic, transmitter supplied flag 0, reduced chi-square 6.68 monostatic rows only (the old workaround) flag 0, reduced chi-square 0.80 So the fit converges on the full set for the first time, where the previous workaround required discarding 18 observations. Aten and Nyx are unchanged controls: neither has bistatic rows, and both move by less than 10%. A residual remains. All of it sits in the 20 bistatic Doppler rows, at about 5 Hz -- a round-trip range-rate error near 0.1 m/s. Ruled out so far: the bounce point (every Golevka delay row is already bp=C, Chesley et al. 2003's re-reduced values), the peak-power reference, station identity and coordinates (DSS-13 to DSS-14 checks out at 21.6 km), the velocity truncation (0.12 Hz, ~1 sigma), and uniform weight inflation (the two tight-sigma rows and the eighteen loose ones agree in Hz, so it is not a weights problem). Closes #528.
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.
Observationcarried one station and used it for both legs, so a bistatic measurement had its up leg evaluated at the antenna that received rather than the one that transmitted. On (6489) Golevka that is roughly 125 sigma per Doppler point -- observed/model 1.039-1.044 on bistatic rows against 0.999934 on monostatic ones -- and 20 of its 40 radar rows are bistatic.Observationnow carries the transmitting antenna's state at the transmit epoch and the up leg uses it directly. Where it is absent the receive station is extrapolated exactly as before, so monostatic behaviour is unchanged.The transmit epoch does not have to be solved for: the round-trip delay is the observable, so it is
t_receive - tau. Golevka's bistatic rows are all Doppler-only and carry no delay of their own, so tau is interpolated from the object's delay rows, which run smoothly from 46.5 to 47.2 s across its 1995 apparition. Evaluating the station rather than extrapolating to it also removes the Taylor truncation, measured at 2.2e-3 m/s in velocity at that round trip.The transmitting antenna is named by
trx, the ADES field, so ADES-derived radar columns need no renaming (stnTxis accepted as an alias).Measured on Golevka, all 40 radar observations:
The fit converges on the full set for the first time; the previous workaround required discarding 18 observations. Aten and Nyx are unchanged controls -- neither has bistatic rows, and both move by under 10%.
One consequence a reviewer should weigh. The C++ gate is
chi2/ndof > 10, so this moves Golevka from flag 2 to flag 0: a loudly-rejected fit becomes an accepted one that still carries the residual below. That is worth stating plainly, but I do not think it argues against merging -- the old behaviour only failed loudly by luck. An object with two bistatic rows instead of twenty would have produced a chi-square under 10 and been quietly corrupted before this change, which is exactly what this issue is titled for.A residual remains, all of it in the 20 bistatic Doppler rows, at about 5 Hz -- a round-trip range-rate error near 0.1 m/s. Eliminated so far, each by measurement: the bounce point (every Golevka delay row is already
bp=C, Chesley et al. 2003's re-reduced values), the peak-power reference (dropping those rows improves only 6.68 to 5.02), station identity and coordinates (DSS-13 to DSS-14 checks out at 21.6 km against a 1.05 m/s velocity difference), the velocity truncation (0.12 Hz, ~1 sigma), and uniform weight inflation (the two tight-sigma rows and the eighteen loose ones agree to ~5 Hz, so it is not the weights).Tests: 5 radar tests pass, on top of the Shapiro work merged in #529 -- the two compose without interference.
Closes #528.