Cell towers/wifi routers not located correctly

Our geo-location of cell towers and wifi routers is not 100% accurate, especially for sparsely populated areas like the Middle of Nowhere.

Most of our data comes from crowd-sourcing, i.e. from lots of people using the app. Readings of signal strength are used to triangulate cell towers and wifi routers.

With time (and with more people using the app) we’ll get a more accurate picture of where the cell towers and wifi routers are.

How can you help us? Using the app a lot, and also using it with GPS. The more precisely you are located the better we can triangulate nearby cell towers and routers using your position.

21 Responses to Cell towers/wifi routers not located correctly

  1. jon says:

    How about a button that I can press that says “I’m currently just below a cell tower”. Or allow me to go to a web site and enter a tower location. Either will be far more accurate than trying to triangulate.

  2. Robert says:

    To expand on jon’s post on 27 March 2011, maybe even email you manually with the cell coordinates?
    Actually, just glancing at your map, in my area, the cell locations aren’t bad. The ones that are obviously not exactly correct, might be off by as much as 300 meters, which isn’t a big deal, for cell coverage. But more accurate locations might be useful, especially in cases where a cell site is on the map currently is further off than a kilometer or so. I just started using this, so I couldn’t say, in my area, if there even are any towers that far off from their real location.

  3. JwM says:

    I don’t see why you don’t just import the actual cell tower location information from the FCC (at least for the USA). Then the tower locations would be accurate within a few feet.
    I drive around western NY quite a bit, trying to get lost, and contributing to OpenSignalMaps in turbo mode, as theres very little detail for my area. I’m on Verizon and my wife on Sprint. We desperately want to plot out a nice coverage map of the area for at least those two networks.

    We tried to visit a few towers that showed up in OpenSignalMaps, but they were either non-existant, or not really a cell tower, which made me wonder where the actual cell tower location information was actually coming from. Plus, I know of two AT&T towers and one Verizon tower that have been in service for over 2 years, and are not on the OSM maps…?

    Can I somehow offer CPU processing power? possibly for my area?

    • Sam says:

      Hi JwM,
      As far as I know the FCC only require you to register a tower if it exceeds 200′, or is within a certain slope from an airport. Looking at the FCC website seems to confirm this – in general they have far fewer towers on their maps than we do on ours. Our data comes from our app which calculates the location of nearby towers and sends it to our servers. Single readings might not be very accurate, but as we get more users and data our averages should improve. Similarly if there are few AT&T/Verizon OpenSignalMaps users in your area we might not have picked up those towers yet.
      If you email me the approximate location of the AT&T/Verizon towers that are missing I can see if there is a specific problem. My email is sam[at]opensignalmaps.
      We don’t have the technology necessary to allow shared processing power, but thank you for the offer!

    • Charles Brandenburg says:

      I like the idea of physically searching out Cell Tower’s, and utilizing our Mobile GPS Technology, to gather exact coordinate’s, and Geo-Location Data. There are a lot of Tower’s in my Area, of Cheney, Washington, 99004, missing, or not listed as well, for my current Cell Carrier, AT&T. I’m a hiker, living in a rural area, with many dead zone’s! Especially, in and around, Federal Reserve’s, like Turnbull National Wildlife Refuge, in Cheney, WA.

  4. JwM says:

    Yes, a “there is a tower right here” button would be nice, but mainly if the person KNEW that it was a cell tower, rather than a random pole that people thought they were magically getting signal from. I’m still trying to figure out how to add locations to google maps from my phone. I was standing at the base of the tower, loading various apps, but I couldnt find one to mark my position, so I could send it on to the opensignalmaps people! grrr!

  5. Allan says:

    It also might be helpful to publish a description about the minimum/optimum data required to reasonably accurately locate a tower. For example, a thousand readings from the identical location probably doesn’t help. With a description of optimum data, some might try and identify data for sparsely covered areas. Also, is there any difference in gathering data for 2G/3G/4G?

  6. Johan Adler says:

    JwM: I’m not sure how much programming it would take to implement it as part of the Android client, but the phone itself has no problem geotagging a photo of the alleged cell tower. Try standing right under the tower, point your camera towards the top of the tower and take a picture (with geotagging activated). The coordinates are embedded in the photo metadata, and the photo itself shows the tower.

    Regarding random poles: Even if it is a cell tower, it may not be the carrier that you are using, so your signal data may not be relevant. On the other hand, one solution would be to let anyone suggest possible tower locations, and then use signal strength data to approximate tower location, if there is a reported tower that fits, use that as tower location unless you find a better fit.

    I guess that it could be useful to get signal readings all around the tower, all 360 degrees, to get the different cell id:s used by the antennas pointing in different directions. I don’t know how far from the tower you would have to be to do that, but I guess that walking in a circle at a distance/radius of a 100 meters from the tower would be fine.

    • Johan Adler says:

      OK, today I park my car close to the closest cell tower to my home, walked up to it and took some pictures. They are available at Picasa using this link and they are all geotagged. The first one might have a less accurate position, as stated in the title.

      After my photo session I drove around the neighbuorhood with your client (still) running at full speed. The cells should be from 240 02 11001 LAC, with cell ID:s 31376, 31375 amongst others. When I am writing this the OpenBmap map is probably not updated with my latest data, but you can see the coverage areas of those two CID:s both having a sharp angle right close to the tower. I tried to get decent pictures that include the actual antenna panels to make it easier to map coverage to panel.

      Is data like this useful? Is it even possible to manually enter a tower location into your database (without spending hundreds of man hours)? Is the location data in the EXIF good enough?

  7. Gary Scott says:

    The main cell tower that serves me is reported incorrectly. It is reported at 32.7595828 -97.1853018 which is in Interstate 30 (in the highway). There is a tower located due east about 3000 feet that has a very strong 1900M signal but I have no idea whether it is Sprint or not. There are so many cell towers all around me its hard to tell, but the one lying east seems to be most likely. The two outliers that support me appear to be fairly accurately located, but not the one that my phone connects to at my house. The CELL CID is 471656 LAC 8035 (Fort Worth, Sprint)

  8. Gary Scott says:

    The other tower near me is located at n32.75856 w97.1792

  9. Kenneth says:

    Here in Denmark all celltowers are in a database called Mastedatabasen, (mastedatabasen.dk). If I compare it to the positioning done by opensignal, all towers all towers a very inaccurate. Even my own router is situated 1 km. away, even though i´m standing next to it. I understand that due to the crowd-sourcing the cell positions will be more precise. Well I have crowd-sourced the same area, and even stood under a known cell tower several times for a half year now, no change in the cell locations has happenend, how´s that. By the way will it be possible for opensignal to read the Cell Broadcast on channel 050. TDC sends out Tower names/positions and cell direction on that channel. I know that this is only done on 2G networks, but when you get the LAC and CID for the 2G, the same LAC and CID goes for the 3G transmitter if mounted on the same mast.

    Kind regards

  10. Rich says:

    I know I’m kind of in the middle of central Virginia but there are only maybe 3 towers in the entire county marked on your map out of tens or more towers in the area. I can drive drive a half hour in any direction and not find a tower on your maps although I never loose signal and it appears I’ve connected to at least a half dozen towers.
    In addition the few towers I do see on your maps are way off, some by miles.
    If you aren’t going to update the towers in Virginia, then your app is rather worthless to me. Too bad, such a nice app *if* there was “real* cell data in your application.

  11. alanjenney says:

    I know that you cannot use the OfCOM Sitefinder Database in the UK as they don’t give the same IDs as you get from the towers, it’s just slightly annoying that I can see my local carrier’s mast and the OpenSignalMaps data shows it a fair way away – about 2000 feet. I could pinpoint it on Google maps in an instant!

    I’ll try to leave background logging on when cycling or driving locally, but cannot promise anything. The tower is in privately owned fields and a distance away from public highways. I cannot get any closer to it to provide more accurate triangulation information!

  12. firemann57 says:

    Overall, this is a really marvelous app. My issue with the cell tower locations is this: On the “Overview” (compass) screen, if I tap the information icon, then the app will report latitude and logitude for a tower, and if I look on Google Earth there is indeed a cell tower located precisely at those coordinates. It is very accurate. BUT when I move to the “Map” screen, the tower locations do not plot at the correct latitude and longitude – even though, as I noted above, the app already has correct knowledge of where the tower is located. Why doesn’t the map correctly plot the app’s very accurate tower coordinate information?

    When I look at tower locations for my service provider on the internet tower map, none of the towers that I connect to are plotted there – even though the app on the phone knows the coordinates of where those towers are.

    I am reporting based on experience in suburban Chicago, Illinois, USA.
    Thanks.

  13. Carlos Martinez says:

    Hi there,

    I would like to know why are you calculating (using the collected data) where the base stations are located instead of using the information provided by operators. They usually report to local authorities where the stations are and those report that information back to citizens.
    As far as I know, it’s been implemented by Spanish Government and by Uk Ofcom.

    Wouldn’t it be more useful to use this information to locate base stations instead of waiting for having enough information of signal strength to triangulate the actual position of the station??

    Thanks and congratulations for your app!

    • gundishy says:

      Agreed, the towers plotted in my area are mostly way off mark and or attributed to the wrong carrier. I like the app and the concept of crowd sourcing signal strength but it would be augmented if the towers were correctly identified.

      The ACMA in Australia has a searchable database of towers.
      http://web.acma.gov.au/pls/radcom/site_proximity.main_page

      and Balint Seeber has already used the database to produce his own superior version using google maps.
      http://maps.spench.net/rf

      If this publicly available data was added to OpenSignal it would almost certainly attract more users in Australia and help build the coverage maps.

      Thanks for the app

  14. ugadog says:

    That’s for this all. I was thinking of actually installing this app, but the reports of the data not getting better leads me to believe crowdsourcing is not being used. What does this crazy app do? It collects a lot of data it seems. How is that used? Reminds me of the SETI program of the ’90s. There is no data on AT&T in Idaho. No towers. It is hard to believe the number 2 provider has no opensignal users but TMobile does. Something is funny here.

  15. Pingback: ► EMF Harassment ◄ - Electromagnetic (Hyper)Sensitivity

Leave a Reply