Thinking about point tracking frequency

[JakeJacobs]

We currently use the breadcrumbs from ArcGIS Mobile for Windows, Collector Classic, and Workforce. Those applications them at a 30 sec interval, regardless of the users' activity.  The devices we use typically have a 3-5m accuracy (just regular iPads and similar consumer grade devices).

We use these breadcrumbs as part of a QAQC process for inspections.  For each inspection location, we look at the gps points for the user who completed the inspection within a time window of when the inspection shows complete (typically the same day, because it's easier).  If one of those points is within a configurable distance (currently using 10m), then the inspection "Passes".  If a gps point can't be found, the inspection "Fails" and the inspection company must do a manual QA to validate the inspection process.

When an inspector is stopped at a location performing an inspection, we get more points over the several minutes they are there. So the relative position of the inspector due to lower accuracy gps isn't too much of a problem.  The gps points get "better" the longer they hold still.  This is good because a)accuracy is lower while they are moving and b)they are often up against buildings when they perform inspections so it can take longer to get a good fix. 

As I understand it, the methodology of Tracker is completely different.  If it detects you are stationary, it won't log any new points.  If you are "walking", it will try to log a point about every 5 meters.

I'm trying to think through our use case and how it will be impacted by switching to Tracker. 

It seems like the Achilles heel is that even if your device gets a "better" location over time from the gps signal when you are stopped at a location, Tracker wouldn't be adding any new points with that better location. Am I missing something that is going on behind the scenes?

I'm also looking for help in thinking about how we would set our parameter for search distance to validate an inspection.  I'm wondering if anyone is using a variable based on the horizontal accuracy of the received points and then perhaps also the activity type.

[Anonymous User]

Hi Erika,

You are correct that Tracker does not request locations when the device is not in motion. Points are collected based on distance and activity (walking=5m, driving=30m, other=30m on iOS).

It sounds like you are essentially doing GPS averaging of the tracks that are collected with Collector/Workforce. This allows you have more precision to identify if the inspector was actually at the location.

This is an interesting scenario. My initial hunch is that because Tracker will be getting locations in the background the entire time the inspector is moving you will not see much change in the precision of the tracks collected between Tracker and Collector/Workforce since the GPS should already be pretty locked in by the time the inspector reaches the location to collect. It seems like something that would need to be verified in the field.

We provide an example script for Workforce that essentially does what you described. It checks when an assignment was completed against the tracks recorded for the worker. It uses distance, time and accuracy as inputs to determine if the assignment was actually completed “on site”. It might be tricky to use activity as sometimes an activity is misclassified (e.g. unknown instead of walking, or cycling instead of driving) but it likely could be done.

[EXOSResearch]

I have a somewhat related question, I think.  I want to use an EOS Arrow device to collect high accuracy tracks (currently trying to use QuickCapture) and export lat-long-timestamp from ArcGIS Online.   Can I do that?