Fixing bug where nmea2tfpose would always assume positive coordinates.

Fixed bug

When using the nmea2tfpose node to produce relative X/Y poses from GGA NMEA messages, all coordinates (regardless of region) were assumed to be positive values. However, coordinate components in S or W directions should be negative in relative X/Y terms.

Steps to reproduce the bug

1. Run the nmea2tfpose node from gnss_localizer.
2. rostopic echo /gnss_pose
3. Publish a GGA NMEA sentence to /nmea_sentence which contains South or West components (e.g. $GPGGA,0,4807.038,N,01131.000,W,1,08,0.9,545.4,M,46.9,M,,*47 - if publishing with rostopic pub, make sure to backslash-escape the $).
4. Publish a GGA NMEA sentence to /nmea_sentence which does not contain South or West components (e.g. $GPGGA,0,4807.038,N,01131.000,E,1,08,0.9,545.4,M,46.9,M,,*47 - if publishing with rostopic pub, make sure to backslash-escape the $).

Expected behavior

The output pose will be different because the coordinates are in different quadrants.

Actual behavior

The coordinates are the same.

Fix applied

If the 4th component of the GGA sentence is S then convert the longitude to a negative value. If the 6th component of the NMEA sentence is W then convert the latitude to a negative value. Same applied for RMC messages except with the correct sentence component positions.

Props to @shan-astuff for figuring this out and implementing the fix on 1.11.0.