Vehicle Position Timestamp and Location Accuracy

[Paul Harrington]

This is a question aimed at feed producers (or those in the know). 

Does the Vehicle Position Timestamp always represent the time at which the included GPS position was taken ?

In other words when the bus tracker sends the latitude and longitude to the server does it also send the timestamp ?

Or is the timestamp later added by a different component ?

What I am trying to ascertain is when you get bus coordinates and a timestamp is there a high level of certainly that the bus was at that location at that time ?

[Stefan de Konink]

On donderdag 25 oktober 2018 13:09:17 CEST, Paul Harrington wrote:
> This is a question aimed at feed producers (or those in the know).
>
> Does the Vehicle Position Timestamp always represent the time
> at which the included GPS position was taken ?

In my opinion this is how it should work. And the timestamp of the entire
message when it started to collect the data.

But some operators are not able to manage this, and always provide the
timestamp of sending.

--
Stefan

[Paul Harrington]

Without knowing the intricacies I would have thought it was fairly basic to send a timestamp with coordinates to a server.

Don't suppose there is a way of knowing which operators do and which don't ?

Or in most cases are the timestamps reliable and it is only the rare operator for whom they aren't ?

[Stefan de Konink]

This is why you would build an integration for a specification operator.
Have some ground engineers checking the data quality case by case.

--
Stefan

[Michael Smith]

While it is not common, I can definitely say that several large CAD/AVL systems do not provide true GPS timestamp. There is no way to know for sure which ones, except by looking very close at the data and looking for anomalies. 

All agencies should definitely require that GPS timestamp be provided when they purchase a tracking system. If the underlying data is not accurate then any system built on top of it will not be as accurate.

Mike

--
You received this message because you are subscribed to the Google Groups "GTFS-realtime" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gtfs-realtim...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gtfs-realtime/cd4f44ce-adca-4cca-a914-a43e71a0c66c%40konink.de.
For more options, visit https://groups.google.com/d/optout.

[David Hudlow]

Yeah, including time with GPS readings is kind of a no-brainer both because it's useful and because time is an inherent part of how GPS works so it's not like it will have a significant impact on complexity.

[Stefan de Konink]

On Thursday, 25 October 2018 22:15:55 CEST, David Hudlow wrote:
> Yeah, including time with GPS readings is kind of a no-brainer
> both because it's useful and because time is an inherent part of
> how GPS works so it's not like it will have a significant impact
> on complexity.

Still I have seen onboard vehicle computers from a German Vendor named INIT
without even synchronizing their Windows clocks to an available GPS signal.
How lazy can you be as system integrator?

I think many of the vendors face the same type of issues, not because their
hardware is bad, but the software is closed source, probably to hide what
is being sold.

--
Stefan

[Michael Smith]

And I have seen database insertion or query time used as the timestamp. The error is of course compounded when the db gets behind by a good number of seconds. 

--
You received this message because you are subscribed to the Google Groups "GTFS-realtime" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gtfs-realtim...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/gtfs-realtime/6928ecec-330d-42b7-9c3f-edef87ff6bbe%40konink.de.

[Paul Harrington]

Thanks Guys, 

Illuminating examples here though I still get the impression that in more cases than not, the timestamps will be good.

Where this is partly coming from is that in various social media (twitter, some FB groups, app reviews) I'm seen a huge amount of vitriol from users over poor predictions.

While I realize some great work is being done (by the likes of MIchael here) to improve predictions, part of me thinks that it needs to be hammered home to users that predictions are just predictions. 

Most commuters are well familiar with their routes (I would think) so once they know where exactly a bus or train is at a given time and know the scheduled time for the 2 stops either side of it, they should be in a good position to judge for themselves when it will arrive. It does however need the GPS position and timestamp to be accurate