Channel Sounding - Speed and Multiple Devices

RE: Channel Sounding - Speed and Multiple Devices

Simonr — Tue, 10 Dec 2024 07:14:55 GMT

1. Not much new I'm afraid. You will get some measurements with 7.5ms, but not enough to do accurate distance measurement. How many samples are needed for an algorithm to be accurate enough is dependent on the algorithm and the accuracy requirement. Do you have an idea of how many samples you'll need for your algorithms? If you don't we can suggest some values, but we need to know what your requirements are is most important.

2. Going through the devices one by one sounds okay, but the effective ACL interval will be higher before you come back around to the same device again here.

There is no support for the LE 2M 2BT PHY on the nRF54L15, that's correct.

Best regards,

Simon

RE: Channel Sounding - Speed and Multiple Devices

MatthewP — Mon, 09 Dec 2024 19:21:12 GMT

Hi Simon,
thank you for your detailed response. Are there any news regarding question 1?

Also, w.r.t to your answer:
1. Since the calculation of the distance is done in the application layer, would it be possible to switch between each device for every ACL interval and aggregate the results on a per-device basis? For example, use 2 ACL intervals for device 1 and save its data, the next 2 ACL intervals for device 2, ... In the meantime, use the data from the steps for each device to calculate the distance more or less simultaneously.
2. Interesting note about scheduling conflicts. Even after reading quite a bit about ACL connection events, intervals, etc. I'm still not a 100% sure how it works. Shouldn't it be fine if I set the procedure count to 1 in the previous example and then go through the connected devices one-by-one?

Additional small question: am I correct in assuming that the nRF54L15 does not support the LE 2M 2BT PHY? I didn't find any mentions in the datasheet.

Thank you again for your great help and best regards,
Matthew

RE: Channel Sounding - Speed and Multiple Devices

Simonr — Fri, 22 Nov 2024 11:24:04 GMT

Hi Matthew

I got a reply from the devs now, and I'm still waiting on a specific answer on question 1, but here are some preliminary responses.

1. This is very situational, but it's worth noting that CS can be started by setting the max_procedure_count=0, that means that CS procedures will continue to run on the ACL indefinitely or until explicitly disabled. In principle you could start a new CS procedure every ACL interval (minimum 7.5ms). 7.5ms is most likely not getting enough CS measurements in that time to do any accurate distance estimatio., but you can configure it so that a new procedure starts once you have collected enough CS measurements in the current CS procedure, to continuously get new measurements.

2. Having multiple peripherals connected at the same time to do sequential CS should be possible in the current NCS 2.8.0 implementation, although we haven't done any testing on this. Note that the central device would be using a lot of memory to store CS measurements from each peripheral device it is connected to and performing CS with. The ACL interval also needs to be noticably higher for the conncetions in order to not get scheduling conflicts between different links.

Best regards,

Simon

RE: Channel Sounding - Speed and Multiple Devices

MatthewP — Thu, 21 Nov 2024 10:32:03 GMT

Hi Simon,

Yes, exactly. The question is basically "under ideal conditions with all devices connected, which ranging update rate can I expect".
Great, thank you!

Thank you for your help, very much appreciated.

BR, Matthew

RE: Channel Sounding - Speed and Multiple Devices

Simonr — Thu, 21 Nov 2024 10:17:41 GMT

As we Channel sounding is still experimental on our end, I'm not sure we have the lowest possible time yet, but under 100ms seems doable at least. I assume you mean a ranging after the devices are connected and synchronized, correct?
I have asked the developers of the channel sounding features in the nRF Connect SDK about this, as I am not sure from just reading the spec and technical overview.
Experimental support was indeed added as of NCS 2.8.0. More details available here.

I'll let you know when I hear back from the devs regarding question 2, and if they have some more input on your first question.

Best regards,

Simon

RE: Channel Sounding - Speed and Multiple Devices

MatthewP — Thu, 21 Nov 2024 05:24:00 GMT

Sorry, my question wasn't clear in that regard and it seems I have missed the support for CS in the release notes for 2.8.0.

So at the moment we aren't testing with DKs at all yet. We wanted to know whether it was even worth trying it out, depending on potential limitations from my first two questions.

In essence we would like to know how fast the positioning can run (in PBR mode for higher accuracy) and whether we can perform the ranging quickly with multiple connected anchors.

RE: Channel Sounding - Speed and Multiple Devices

msip — Wed, 20 Nov 2024 20:14:15 GMT

Are you testing between with two nRF54L15-DKs using nRFConnect SDK 2.8.0?