Using compass in smartphones to improve accuracy of GPS reading

#1

Hello,

Has anyone worked on an app or a solution to combine GPS reading from a smartphone with data from the compass? This would be helpful to improve GPS reading accuracy in assessing buildings in dense urban areas.
Grateful for any experience sharing or suggestions!
Fabrizio

#2

Interesting :slight_smile: Is the goal to assist in determining the device’s coordinates or to report the compass heading of the device? I haven’t heard of a way for the compass to help the GPS acquire a fix, but some devices may already use Assisted GPS or OS-specific geolocation techniques, like scanning nearby WiFi SSIDs. Probably the next step would be to look at the APIs provided in HTML5 (for cross-platform Enketo) and Android (for Collect) to see what sources they use for location data and how those may vary from device to device.

#3

Thank you @jnm
We are trying to reduce inaccuracies of GPS reading when identifying a building when damaged by a disaster. We take the GPS reading and a picture of the building, the metadata from the picture has also the GPS coordinates, we would like to add metadata from the compass so that we know which building was assessed by triangulating this info. In dense urban areas the GPS reading may not be accurate or cover more than one building.

#4

Ah, that makes more sense now (strictly speaking you cant use a compass to improve either the accuracy or acquisition time of a GPS fix).

What you could do is, in addition to your photo + geopoint question, add a bearing question to your form and inform the user to point the device in the direction of your desired target. If you do this looking at a particular target from a couple different locations, you could conceivably do a little triangulation to get an approx fix on the target itself (assuming you cant physically get on top of the actual target to capture a fix).

#5

Thank you very much @Xiphware, we will test that!
Apologies for getting back to you only now, I have been offline for few days.

#6

YW. I’m very curious to know how well a triangulated GPS ‘fix’ you can get with this approach, so please report back with any field results when you have them! Cheers :slight_smile: