search

openbmap / radiocells-scanner-android

WLAN and cell tower scanner for Radiocells.org
https://www.radiocells.org
Other
55 stars 26 forks source link

mobile car hotspots #180

Open mega-stoffel opened 7 years ago

mega-stoffel commented 7 years ago

Hello all! A few weeks ago I discovered this great project and I'm collecting all WIFIs and cells I can find in Stuttgart. What I found out: There are quite a few cars with mobile hotspots. These are named e.g.: audi_mmi_XXXX mb wlan XXXX audiXXXXX seat_wlan_XXXXXX mb hotspot XXXXX

That these are car hotspots is just a guess, depending on their names and that I can measure them on my bike "longer" than regular, stationary WIFIs. Is there a way to verify my guess?

If so: I would avoid to add these mobile hotspots to the database. Would it be possible to add these names to the blacklist or would it be even better to use the MAC address for blacklisting e.g. all network adapters from Audi? If Audi (in this example) is the owner of a specific MAC address range at all?

wish7code commented 7 years ago

Hey! Glad you like our project! Feel welcomed :-)

Would it be possible to add these names to the blacklist or would it be even better to use the MAC address for blacklisting e.g. all network adapters from Audi? If Audi (in this example) is the owner of a specific MAC address range at all?

I can confirm from my own observations the wifis you mention are build-in car wifi systems. For the scanner in fact we've got some MAC vendor ID filtering in place, which filters out some of these based on the wifi manufactor id. Wifi manufactor for most German car brands is Harman International Industries if I remember correctly. Neverthless we haven't updated the lists since a while, so we don't catch all mobile car networks.

But technically MAC based filtering might not be the best approach: For example while riding the highway I often see truck drivers have their 'home-brew' car wifi in place , i.e. I see a lot of Huawei mobile networks or even Fritzboxes moving along the road. One wouldn't be able to catch these with a pure MAC based filtering. @mvglasow then suggested not to rely on MAC filtering, but solve this issue server-side, by ignoring wifis which have been seen in different locations. For a detailled description see #47

I still like this proposal (or better: a combination of both), but haven't found the time to implement it yet.

mega-stoffel commented 7 years ago

just today I drove in the local S-Bahn, which has WIFI on board. Therefore I had 90 measurements of a very high quality, recorded over many 100m (I also took a screenshot, if you'd like to see it). I think, these kind of measurements should be recognized by the client as "moving targets" and should not be uploaded. It might be easier to realize in the app some kind of moving targets (a stationary WIFI should never be visible over such a long time and distance with an almost constant signal strength) than in any database post-processing function.

TRSx80 commented 6 years ago

I found it very interesting how DejaVu is handling this problem: https://github.com/n76/DejaVu#moved-rf-emitter-handling