BICAS: Use L1/L1R QUALITY_BITMASK for detecting BIAS sweeps

[ErikPGJ]

Context

BICAS uses knowledge of BIAS sweeps for removing data from L2 datasets.

Sweeps are currently identified using "hacks":

• SBDA (Sweep Bias Detection Algorithm): Before a hardcoded time stamp, mux mode=4 is interpreted as a sweep.
• SCDA (Sweep Current Detection Algoritm): After a hardcoded time stamps (above), BICAS searches for time windows where the min-max difference for HK bias currents (during mux mode=4) exceed a hardcoded values and interpret them as sweeps.

TODO

The intended long-term solution is to identify sweeps using L1/L1R QUALITY_BITMASK quality bits. These are not available in production by ROC as of 2024-11-29.

[ErikPGJ]

The L1 samples generated by updated ROC s/w, Xavier Bonnin, e-mail 2024-11-25, support sweep quality bits but are buggy.
https://rpw.lesia.obspm.fr/roc/data/private/users/xbonnin/outgoing/l1_cdf39_test_datapack/L1/

I looked at some of your L1 sample files to see if they now support the quality bit for thruster firing (bit7=offset 128) but 
as far as I can see they do not. I looked at a few days with known thruster firings (visible in data) and to me 
QUALITY_BITMASK seems inconsistent and buggy unfortunately and also fails for sweeps sometimes (effectively both 
false positives and false negatives). Is this correct?


solo_L1_rpw-lfr-surv-cwf-cdag_20230401_V01.cdf: Contains superficially realistic QUALITY_BITMASK values, incl. s/c roll 
(but nothing for thruster firings). Sweep is labelled by bits.

solo_L1_rpw-lfr-surv-cwf-cdag_20230403_V01.cdf: Contains only QUALITY_BITMASK=255 which I believe used to be the 
old value (I assume a faulty fill value). Must be wrong because the entire day is not thruster firing or sweep anyway 
(actual sweep at end).

solo_L1_rpw-lfr-surv-cwf-cdag_20230406_V01.cdf: Contains only QUALITY_BITMASK=0, despite thruster firing which I am 
guessing implies a s/c roll(?!). Actual sweep at end of day is not identified.

solo_L1_rpw-lfr-surv-cwf-cdag_20230409_V01.cdf: Contains superficially realistic QUALITY_BITMASK values, incl. s/c roll 
(but zero for thruster firings). Actual sweep correctly identified.

solo_L1_rpw-lfr-surv-cwf-cdag_20230411_V01.cdf: Contains superficially realistic QUALITY_BITMASK values, incl. s/c roll 
(but zero for thruster firings). Actual sweep correctly identified.

solo_L1_rpw-lfr-surv-cwf-cdag_20230414_V01.cdf: Contains only QUALITY_BITMASK=0, despite thruster firing which I am 
guessing implies a s/c roll(?!). Actual sweep in middle of day is not identified.

/@ErikPGJ e-mail 2024-11-29

[ErikPGJ]

L1/L1R QUALITY_BITMASK bits (one per ANT) for sweeps are available, but they cover way too much time, or possibly I am confusing it with the quality bit for thruster firings.