Occasional null values for GNSS Metadata (iOS/Trimble R2)

[Anonymous User]

Every once in a while, we are getting point data in with null values for metadata. It's really infrequent, but I cannot figure out why it is happening. The geometry gets collected, and the associated attributes are recorded, but the accuracy information is not present. Like I said, it's less than 1 in 10 points that do this, but it's definitely something we need to figure out. 

We're on Collector 10.4.1, iOS, Trimble R2 using GNSS Status app, and we're getting internet based RTK corrections. 

[KevinBurke]

Hi Brook,

Thank you for sharing that information. I recently tried to reproduce this issue using an iPad Pro (iOS 10.2) and a Trimble R2 with and w/out RTK corrections and I was unable to reproduce the issue. For my feature service, I had enabled all of the GNSS metadata fields per the online documentation. My testing lasted for over 1 hour and approximately 80-100 points were captured and all the metadata was populated. My colleague also tested with a Trimble R1 and an iPhone 7 and couldn't reproduce the issue either.

Could you provide more details of your testing?

- What type of iOS device(s) are you using?

- When you do reproduce the issue, is it just a particular GNSS metadata field that's blank or is it all the fields?

- If its only a single field, is it the same field all the time or can it vary?

- In your feature service, which GNSS metadata fields did you include? All of them per the documentation or just a few?

Thank you for that information.

-Kevin

[Anonymous User]

Hello Kevin,

We're using the iPad Pro and iPad Air 2. This is happening on both devices, on iOS 10.2 running Collector 10.4.1 and Collector 10.4.2. 

Also, to answer your question, it leaves all metadata fields null. 

I've included all metadata fields in the manner first described back in June. At the time, I even posted a tbx for everyone before the python scripts were published: Add necessary GNSS fields for Collector App metadata logging 

I've read over the final documentation: Record GPS metadata—Collector for ArcGIS | ArcGIS 

I'm fairly certain no one is manually creating geometry. And, we have an accuracy tolerance set, so the user shouldn't even be able to update if "GPS metadata information is not available for the GPS position". 

Is it possible that the metadata expires? I know the location information is cached at the moment you select the feature type to be collected, and the cache is applied to the DB when you hit update. I guess it is possible they could be waiting too long to hit update...?

I appreciate you looking into this. 

Brook

[KevinBurke · Accepted solution]

Hi Brook,

Thank you for that information. What I suspect might be happening here is that when the field user clicks the 'Submit' button in the upper right corner, they're inadvertently tapping on the map itself, thereby creating a map point and no longer a GPS point which will clear the GNSS metadata fields of their contents. For Collector on iOS, when working with tablets there is a small gap between the top of the collect panel and the bottom of header. Sometimes when going to tap on 'Submit' if you tap too close to that gap, it will actually drop a new point there. This is a current design limitation that's being looked into for a future release.

The reason why I suspect that's what's happening is for a few different reasons; 1) this behavior happens intermittently and is not easily reproducible, and 2) when this issue happens all of the metadata fields are blank as oppose to just a few of them. Several months back, my colleague and I saw an issue where random metadata fields would be blank after having captured several points. Not all of them would be blank, only a few. But those few that were blank would fluctuate between different fields. This issue has since been addressed already and we haven't been able to reproduce it since then. Therefore I don't suspect that's what's happening here.

I have a question for you. For the points that have the blank GNSS metadata fields, when looking at the map is the geometry in the correct locations or is it off? If you are running into the issue where the field user is tapping on the map when hitting 'Submit', the offset of that point should be noticeable.

About how many field users/devices do you have?

To answer your question, no, the metadata values that are already written into the fields when the GPS point is dropped on the map doesn't expire. The time it takes the user to actually submit the feature after the point has been added to the map, shouldn't have any bearing on this behavior.

Thank you.

-Kevin

[Anonymous User]

Kevin, 

I'll leave the draft of my reply to you below for posterity, but it turns out one of our users was manually inputting hydrant assemblies... It seems some of our other users were accidentally placing points in the map with their finger. 

Thank you for the quick reply. I am almost certain the issue is not from accidentally placing the point with their finger. The locations of these points seem to be spot on. We have hydrant assemblies which contain a fitting along the main, a 90 degree bend, a valve, and then a hydrant. All of these four points are in the location you would expect, and they are all within an appropriate distance from each other. 

One thing I noticed just now was these points seem to be grouped together spatially (which would also mean temporally). The user collected an entire street's worth of data, but somewhere in the middle, the metadata temporarily stopped getting captured. The one instance I'm speaking of is an entire hydrant assembly. All four points (across four different feature types) are missing metadata. The points taken before and after, on the same device, by the same person, have metadata.

It seems to me the user would have to go through a lot of effort to be manually putting these points in after the fact. 

Apparently I was wrong.

Thanks so much for your time. It's good to know this product is so reliable. 

Brook 

[Anonymous User]

I've been thinking maybe I should make one of the metadata fields a requirement in the DB. That way, we would prevent users from doing this. I'm kind of worried about that type of practice though.

Do you plan on putting in options to prevent users from manually placing points?

Thanks again,

Brook

[KevinBurke]

Hi Brook,

Thank you for that information. In general, with regards to having a setting that only allows features to be captured via the GPS and not tapping on the map, this is something we're aware of and are planning to implement in the future. Unfortunately I don't have any specific details at this time as to when/how that functionality will be included.

With regards to the proposed workaround for making the GNSS metadata fields required in order to force the field user to capture data via the GPS, in theory this should work just fine because of the fact that by-design the GNSS metadata fields are hidden when editing and capturing new features. Therefore if the user has a stable fix and capturing a point, the field(s) should be populated on the back-end automatically thereby making the 'Submit' button active in order to submit the feature. However, this could also lead to confusion among the field users because if they do happen to capture a point by tapping on the map, the 'Submit' button will remain grayed out even if they've populated all the attribute fields. The reason being is because the hidden metadata fields will not be populated and will prevent the feature from being submitted. The user will not be notified as to the reason why. It is therefore not recommended to use this workflow since GNSS metadata fields weren't designed to be made required, because they get populated automatically and require no user input.

- Is there a particular reason why the field users are capturing points with a combination of GPS and manually tapping on the map?

- Are they physically in areas where they cannot obtain a GPS fix with the receiver and are therefore just adding points manually?

Thank you.

-Kevin

[Anonymous User]

Hello Kevin,

Thanks for the heads up about the future functionality. I kind of assumed that would be the case.

As for your questions about our users: No one is supposed to be manually placing points in the map. We have R2, Sub-Foot, Multi-Constellation, with access to a local RTK network. We can get 6" accuracy while sprinting under a row of Live Oaks 🙂 For some of the guys it was clearly a mistake of them accidentally doing it. For the one guy drawing hydrant assemblies, maybe he convinced himself that since they are so standard, and they are partially above ground, getting good accurate data for them was not as important as saving his own energy... I'm pretty sure he won't be doing that anymore.  

Thanks again for your help and assurance. Looking forward to these improvements. 

Brook