Using OpenCellID Database for Position Lookup if no GPS Avaiable

I am afraid the accuracy of CellID localisation is too low. At the same time the usage of this feature will be also low, as it needs specific approach around trip detection and fuel consumption, It is interesting to have it for a small segment of cargo/personal/auto trackers, in case it is lost an gps blocked. But most of these trackers can send that data on demand by sms commands, and you can manually find the location.
Anyway, as an app it would be handy sometimes.

Agree Wolf this is a good back up if the GPS is lost. We have all of our BTS locations maped as POI's as we can get them from our service provider. We have a deal with them that we give them signal strength information for them to be able to get user data on the network operation and we gate location information. It has my vote

Regards
John

Hi Aliaksandr,

great to hear that!!!

Now we need a Hardware for that hihi
(I know that Jamal developed something like that)

We are also interested in LBS support by wialon software

Hi Shal,

sounds great.

An idea would be to do some kind of calculations...
For example in larger towns you have typically every 500m a new BTS (base station transceiver). So you could make a circle of 250 meter around the LBS position.
Outside the city there are typicaly every 3-6km a BTS. so here radius should be round about 2KM.

Idea for understanding how big the circle should be:
do a calculation of how many BTS you have per sqarekilometer (if you use opencellid and/or other sources you can just do a grid and than have this calculation done once....)

Wolf.

Wolf, as far as the appropriate hardware to use for this purpose, you might want to look into GoSafe. I know their devices can capture and transmit:
MNC, MCC, LAC,  Cell Tower ID and RSSI (tower signal strength). They can send data for up to seven cell towers at a time.

Hi Shal,

we are going to receive a brand new low energy using GSM tracker which we would like to implement into Wialon. So the question would be, how far you are with the integration of LBS (Celltower) Data into Wialon?
Is there any chance to speed up that?

Thanks a lot
Wolf.

Hi Shal,

playing around with opencellid-database and comparing to google maps cell lookups here are some screenshots of triangulation by doing simple weighting of 3-5 celltower positions. weighting is at the moment done by using the position in table of neigbhourcells and not using available TA parameter and/or signal-strength... (i was analysing the data of our little new HW and signal strength is not always a good indicator, the lookuptable of the neighbourcell-list seems to be already ordered by the gsm modem in a good way)...

the first shows a cell-list of a device on my office-desk (no gps)... calculated distance ca. 200m. distance of cells is only a few hundred meters.

second shows a testunit in denmark near Ejsberg. Cell-Tower Distance is here much higher (up to 7km) and the maximum difference in elevation is 1m or so :-)) so gsm modem receives also cell-towers with longer distance... So calculated location is round about 1000m away from real location.

going to compare it with google data :-)

Wolf.

P.S. Wialon is also a nice mapping-tool! together with POI and Geofences hihi

The LBS function will return approx coordinate of cell tower. The idea is to have smth, when there is no GPS. Initially it will be user function, not a system - e.g. click on asset and do LBS navigation detection instead storing its coordinate in its history like for GPS.

Hi Globalsoft,

we are just developing our own little "LBS" Translator (already working but still in debug/stability phase).
It receives the data from the hardware we use (little device with battery inside, gps inside, active RFID tag inside, movement sensor inside, light sensor inside), if no gps available it sends up to 10 celltower ids together with TA, if it can not connect to server it starts sending via RFID its ID and you can use a special antenna to find the unit (even inhouse) It uses two Goldcaps to use 100% of battery energy and has a (tested) stand by time of 3 years (or 15 years) at 2 messages per day.
(my testdevices sending every 15 minutes and already sended 980 messages in 4 weeks and battery is 90%)

our little server is installed at a windows virtual server and has the opencellid.org cell-list on a mysql database (4 Million cellids with location).
If a new message by UDP is received from server, it scans the message and does a lookup of celltowers in the local database. than it calculates the different distances and tries to do some kind of weighting. than it uses Wialon IPS protocol and sends the calculated "position" to Wialon (together with some infos we receive from device). Additional the little server logs into sdk.wialon and sets a geofence with the calculated radius (so if gps=valid, radius is a few meters, if GSM it uses the TA (timing advance value) of the served cell to create the radius) to the middle of the served cell and you can see it in Wialon.
In a future version we want also to show the served cellid with location as POI and dynamicaly change / update / delete them

Usage for this HW is always anti theft where there is no regular power suply and parcel tracking

Wolf

Hi Shal,

here is a case for your debug guys :-)
Device "GORDON" = VT310, it tells for MCC=106 but that is hexadecimal.
(rest of data is also hex as i remember), so that does not work with LBS lookup...
I know that some HW sends data as HEX and some as decimal, so how Wialon handle that?

Wolf.

Hi Vestigodata,

seems that there are data missing:

you need MCC (mobile country code), 3 digits
MNC = mobile network code, 2-3 digits (seems to be 241???)
Area code (or location area), cellid

so one is missing

Wolf.