Indoor Positioning possible / planed?

Location
#1

Hi there, in private I contribute for years to Freifunk (free wifi operrator) and OpenStreetMap (crowdsourced geodata). So I have some interests in wifi spectrum and geoprocessing :slight_smile: And yes I followed and contributed to MLS for quit some time …

In my day job I work at the IT departement of a small university with a complex building. So I started to create an floor plan based on OpenStreetMap, but would love to see users being able to estimate their 3D position within our building.

Unfortunatly for indoor there seem to be no crowdsourcing wifi wardriving project beacause of the lack of GPS. And there are only some limited appraches like c3nav, find3, … that work on a dedicated campus and require specialised apps.

So I like to ask, if Mozilla has any plans to work on indoor locations?

#2

Thank you for reaching out @HMT-HRO-MaMe!

So I like to ask, if Mozilla has any plans to work on indoor locations?

There are no plans to work on indoor locations.

The API allows submitting Bluetooth Beacons along with a location, and I think the intention was to support indoor location finding. However, Bluetooth observations are around 170,000, versus 59,000,000 for cell towers and 2,400,000,000 for WiFi. I don’t think our Bluetooth observations are a sufficient database for general indoor location determination.

#3

Hi, thank you for the feedback.

Beside the radio aspect, there seems currently a limitation to use GPS / GNSS as location reference to track the position while recording the samples? So we can’t record anything in buildings (without GPS coverage) or on different floors, right?

Over at OSM there are some years of experience in indoor modelling and maintaining this data. For example, we added our university: https://indoorequal.org/#map=18.91/54.091616/12.1444578&level=0
But of course the value of the data would be optimal, if users could estimate their position. And in best case without installing a dedicated application, but with support from your mobile browser :slight_smile:

As univeristies provide bigger campus wifi systems with dozens of accesspoints, it should be a perfect scenario to estimate the location. What if the community would share the AP MAC adresses and positions via OSM? Could this be integrated to MLS or is this a to different data model without any measurement samples?

#4

The MLS data model is based on observations, not on device positions. See https://ichnaea.readthedocs.io/en/latest/algo/observations.html for details and diagrams.

To get the data into MLS, you would submit the WiFi MAC addresses (checking that the SSID does not end in “_nomap”, see https://location.services.mozilla.com/optout), their relative strengths, and your position from OSM, with a source of manual or fused. See the API doc for details.

Getting the mobile browser to use MLS as the location source is a much harder problem. Android devices with Google Play Store support probably fall back to Google Maps geolocation. Apple devices use Apple geolocation. See https://wiki.mozilla.org/CloudServices/Location/Software for more info (could be out-of-date).