BeiDou antenna patterns ignored in conversion from ANTEX

[AndreHauschild]

Description

The current igs20.atx contains multiple records for three BeiDou-3 satellites (C215, C226, C225) with different PCO values and different validity periods. The program GnssAntex2AntennaDefinition ignores reoccurring entries for the same SVN because a time-dependency of of antenna patterns and offsets is not foreseen in the antenna definition files. The corresponding warnings are:

antenna 'BEIDOU-3M-SECM.C215.2018-078A' already in the list with different values -> skipped
antenna 'BEIDOU-3M-SECM.C215.2018-078A' already in the list with different values -> skipped
antenna 'BEIDOU-3M-SECM.C215.2018-078A' already in the list with different values -> skipped
antenna 'BEIDOU-3M-SECM.C226.2019-078B' already in the list with different values -> skipped
antenna 'BEIDOU-3M-SECM.C225.2019-078A' already in the list with different values -> skipped
antenna 'BEIDOU-3M-SECM.C225.2019-078A' already in the list with different values -> skipped

As far as I see, a revision of the antenna definition format would be necessary to properly handle this.

GROOPS version

main (latest commit)

Operating systems

• Linux
• Windows
• macOS
• Other (please specify in the description)

Log output

[Hanbing59]

Hi Andre,

Interesting! I did not check the code yet to see how it identify "reoccurring entries". As far as I know, in ANTEX files, two (or more) entries for an physically identical antenna can co-exist by different validation periods, for example, because of de-commission and re-commission events. This is not a BEIDOU-specific phenomenon.

So I find it interesting if it can handle this situation well for other satellites (or more generally, antennas) but not for BEIDOU satellites with different PCO/PCV entries.

Regards,
Hanbing

[Hanbing59]

Indeed, it seems that the valid time period attribute of patterns of antennas haven't been considered.

In GnssAntex2AntennaDefinition::addAntenna, it checks the name, serial no., radome, no. of patterns and for each pattern, the type, resolution and values to filter out duplicated antennas and report in-consistent records for the same antenna. Without the valid time attribute attached to each pattern, antennas with different patterns for different time period would be regarded as "in-consistent entry" and therefore, be skipped.

In the case that antennas have multiple entries sharing all the same info (including pattern values) but different valid periods, it would just be regarded as duplicated records and be filtered out.

So, yes. I think @AndreHauschild is right. Maybe an additional data attribute should be added to the class GnssAntennaPattern and some following modifications are needed.

[tmayerguerr]

Hi Andre,

The issue should now be resolved. Same SVNs with different PCOs/PCVs are now distinguished based on the radome (which contains the COSPAR ID for satellites). As before, the time mapping is handled via the TransmitterInfo file. I hope I haven't created any new problems by doing this.

Best regards
Torsten