• State Verified Answer
  • Locked Locked
  • Replies 11 replies
  • Subscribers 63 subscribers
  • Views 3510 views
  • Users 0 members are here
Attachments (10) Download All
Nordic Case Info
  • Case ID: 276552
Options
This post is older than 2 years and might not be relevant anymore
More Info: Consider searching for newer posts
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Multicell Geolocation via LTE

jdorn

Hello,

I am attempting to implement cell tower triangulation using unwiredlabs.com's api.

I have single-cell geolocation working. I am having trouble getting multi-cell geolocation working.

I have two issues with my current implementation:

1) lte_lc_neighbor_cell_measurement() trtiggers an LTE_LC_EVT_NEIGHBOR_CELL_MEAS event, but that event does not always provide neighbor cells. In fact, it rarely does so. Is there a way to make the system return at least a few neighbor cells consistently other than when there is a bad signal?

2) The neighbor cell measurements event only provides the physical cell id and the signal strength. I want the LAC/TAC and normal cell id of the towers as well. Is there a way to get this information? 

I have examined half a dozen other related devzone threads (linked at the end), but they are all older than the latest modem firmware and the multicell location example so I was hoping that the situation has changed.

If the above two items are not possible, could I disconnect from my current cell tower and force a connection to a different tower? If so, how can I do this?

I am using the nrf9160dk rev 1.00 with modem firmware 1.3.0.

https://devzone.nordicsemi.com/f/nordic-q-a/73139/scanning-for-all-nearby-base-stations-with-nrf9160dk

https://devzone.nordicsemi.com/f/nordic-q-a/66777/can-i-scan-more-cells-than-one/274337#274337

https://devzone.nordicsemi.com/f/nordic-q-a/44093/nrf9160-dk-geolocation-through-lte/173492#173492

https://devzone.nordicsemi.com/f/nordic-q-a/74267/nrf91-geolocation-and-cell-triangulation/306129#306129

https://devzone.nordicsemi.com/f/nordic-q-a/53189/nrf9160-cell-tower-information-for-geolocation-with-lte-and-cell-tower-id-s-limited-nbrgrsrp/214805#214805

https://devzone.nordicsemi.com/f/nordic-q-a/42310/nrf9160-zephyr-ncs-zte-m-nb-iot-multiple-cell-tower-rssi

Parents
  • +1

    Hello Jeremy,

    jdorn said:
    Do you have any updates on how to improve the reliability of the neighbor measurements? If this is something where I need to share code, I can create a private query and do it there.

    my apologies for the late reply. It took us some time to analyse your traces.

    jdorn said:
    So I figured out that issue, it was a bug on my end. However, I am still inconsistently able to find neighbors. Anecdotally, the find rate seems to have improved from 1 in 10 or 20 to 1 in 3 or 5. Traces for scenarios where neighbors are found and not found are included.

    To start with, cell mapper locates cells recorded in this log to the middle of Chicago centrum area.

    trace-neighbor-found:

    It seems that we are on rather strong cell on band 12 in T-Mobile network. Tower is enb id 20934. It has both cell id's 32 and 453. However, all the neighbours which are close to this tower seems to be on different bands (2,41,66). Inter frequency measurements are not active, so these bands are not searched.

    trace-no-neighbor-found:

    We seem to be on AT&T network at band 12 and there are even less intra neighbours available. The signal level is as high as in T-Mobile network. Tower is enb id 210917.

    So the main point here seems to be that the band 12 cells are the best for both T-mobile and AT&T, but apparently there are not too many neighbours on that band/frequency i.e. intra frequency. And the inter frequency neighbour measurements are not active. If you would like to see more cells, maybe bandlock could be used to exclude band 12 or simply allow only 2 and/or 66? Risk is that 2/66 doesn't have Cat-M cells.

    To summarise this: neighbour cells measurement service in current modem release is implemented so that there would not be significant increase in power consumption in any scenario. It can detect any LTE Cat-M neighbour cell on the same frequency, or LTE NB-IoT neighbour cell when operating on LTE NB-IoT, but detecting inter-frequency neighbours requires that network is configured to provide information about those neighbours to devices. In this case network has not been configured to do that.

    But we are continuously working on improving our software. The %NCELLMEAS functionality was recently introduced with modem firmware 1.3.0, and new features will be implemented in future releases.

    I hope this will help you!

    Regards,

    Markus

Reply
  • +1

    Hello Jeremy,

    jdorn said:
    Do you have any updates on how to improve the reliability of the neighbor measurements? If this is something where I need to share code, I can create a private query and do it there.

    my apologies for the late reply. It took us some time to analyse your traces.

    jdorn said:
    So I figured out that issue, it was a bug on my end. However, I am still inconsistently able to find neighbors. Anecdotally, the find rate seems to have improved from 1 in 10 or 20 to 1 in 3 or 5. Traces for scenarios where neighbors are found and not found are included.

    To start with, cell mapper locates cells recorded in this log to the middle of Chicago centrum area.

    trace-neighbor-found:

    It seems that we are on rather strong cell on band 12 in T-Mobile network. Tower is enb id 20934. It has both cell id's 32 and 453. However, all the neighbours which are close to this tower seems to be on different bands (2,41,66). Inter frequency measurements are not active, so these bands are not searched.

    trace-no-neighbor-found:

    We seem to be on AT&T network at band 12 and there are even less intra neighbours available. The signal level is as high as in T-Mobile network. Tower is enb id 210917.

    So the main point here seems to be that the band 12 cells are the best for both T-mobile and AT&T, but apparently there are not too many neighbours on that band/frequency i.e. intra frequency. And the inter frequency neighbour measurements are not active. If you would like to see more cells, maybe bandlock could be used to exclude band 12 or simply allow only 2 and/or 66? Risk is that 2/66 doesn't have Cat-M cells.

    To summarise this: neighbour cells measurement service in current modem release is implemented so that there would not be significant increase in power consumption in any scenario. It can detect any LTE Cat-M neighbour cell on the same frequency, or LTE NB-IoT neighbour cell when operating on LTE NB-IoT, but detecting inter-frequency neighbours requires that network is configured to provide information about those neighbours to devices. In this case network has not been configured to do that.

    But we are continuously working on improving our software. The %NCELLMEAS functionality was recently introduced with modem firmware 1.3.0, and new features will be implemented in future releases.

    I hope this will help you!

    Regards,

    Markus

Children
No Data
Related

core_v2_widget.Hide()