Ticket #431 (new defect)
ogpsd mangles timestamp, doesn't allow sub-second sample frequencies
| Reported by: | vnevoa | Owned by: | daniel |
|---|---|---|---|
| Priority: | major | Milestone: | |
| Component: | framework/ogpsd | Version: | |
| Keywords: | timestamp precision | Cc: |
Description
When using non-fso based distros, I've used the GTA02 many times as a high speed GPS logger for research projects. I would just send the "CFG-RATE" u-blox messages into the chip to configure it to output 4 NMEA samples per second and eveything just worked.
Now that I'm using SHR-unstable with FSO underneath, this doesn't work because the output of "gpspipe -r" only reports 1 sample per second.
Investigating this problem, I saw that it is the "ogpsd" framework component that sends it's updates into DBUS with a timestamp field of an integer number of seconds. And it does send 4 updates per second (just like I configured in the chip), but in this case it sends each set of 4 updates with the same exact timestamp. This confuses "fso-gpsd", which filters out 3 of the 4 samples per second.
Please allow for a hundredth of a second resolution in the timestamp to allow more than one sample per second (just like the GPS chips do).
Attached you will find my python client that dumps the "PositionChanged?" events and an example of the output.
To configure the chip for 4Hz sampling rate, you may include the following line in "/usr/lib/python2.6/site-packages/framework/subsystems/ogpsd/om.py", function "initializeDevice": self.send("CFG-RATE", 6, {"Meas" : 0x00fa, "Nav" : 0x0001, "Time" : 0x0000})
Thank you!
Attachments
Change History
comment:1 Changed 3 years ago by stefan
Daniel, how does allowing a higher resolution for timestamps sound to you?
comment:2 Changed 2 years ago by sim
decoration Changed 1 year ago by admin
bathtub Changed 1 year ago by admin
solar system Changed 1 year ago by admin
stair parts Changed 1 year ago by admin
solar supply Changed 1 year ago by admin

