BLE Channel Sounding - changing the number of samples.

RE: BLE Channel Sounding - changing the number of samples.

Vinay Bhagure — Fri, 24 Jan 2025 08:16:26 GMT

Thankyou Edvin!

RE: BLE Channel Sounding - changing the number of samples.

Edvin — Thu, 23 Jan 2025 23:25:48 GMT

In this sample, you would have to do that in your application. I guess it is just some simple maths, but you need to decide how you want to weight them. If you weight them equally, just add together all measurements times the number of samples used, and divide by total samples.

But again, neither of these are good measurements at the moment. My best advice is to wait for the sample to get some smart algorithms that actually gives decent results.

Best regards,

Edvin

RE: BLE Channel Sounding - changing the number of samples.

Vinay Bhagure — Tue, 21 Jan 2025 12:39:56 GMT

Is there any way to combine both the distances from RTT and PBR methods and get the average solution for the distance measurement?

RE: BLE Channel Sounding - changing the number of samples.

Edvin — Wed, 15 Jan 2025 10:24:50 GMT

Indeed this is not the final solution for the channel sounding. It is currently just a proof of concept, as I mentioned, to show that we can fetch the raw data correctly.

Best regards,

Edvin

RE: BLE Channel Sounding - changing the number of samples.

Vinay Bhagure — Mon, 13 Jan 2025 10:01:11 GMT

Thankyou. That was exactly what I was looking for. But event if we change the number of samples, there is inconsistant and inaccurate values from RTT and PBR methods. It might take some time for Nordic to develop a consistant working solution for this.

RE: BLE Channel Sounding - changing the number of samples.

Erik Sandgren — Fri, 10 Jan 2025 14:51:26 GMT

Hello,

Not sure if this is what you are looking for, but one way of changing the ratio of main mode to sub mode steps is by changing these two values https://github.com/nrfconnect/sdk-nrf/blob/509fa66004370f00faf13912cc6fced25dda6d6c/samples/bluetooth/channel_sounding_ras_initiator/src/main.c#L423-L424
For example if I change both to 1, I get roughly as many sub mode steps (RTT steps in this case) as I get main mode steps (PBR steps in this case). Example log from the sample with this change:

I: Estimated distance to reflector:
I: - Round-Trip Timing method: 4.592482 meters (derived from 78 samples)
I: - Phase-Based Ranging method: 0.434500 meters (derived from 79 samples)

RE: BLE Channel Sounding - changing the number of samples.

Edvin — Fri, 10 Jan 2025 13:20:10 GMT

Hello,

Did you test the sample?

Running it for a short time gives me this log:

I: - Round-Trip Timing method: 2.443308 meters (derived from 10 samples)
I: - Phase-Based Ranging method: 1.531434 meters (derived from 147 samples)
I: Subevent result callback 1
I: Ranging data ready 1
I: Ranging data get completed for ranging counter 1
I: Estimated distance to reflector:
I: - Round-Trip Timing method: 3.147821 meters (derived from 9 samples)
I: - Phase-Based Ranging method: 1.485374 meters (derived from 148 samples)
I: Subevent result callback 2
I: Ranging data ready 2
I: Ranging data get completed for ranging counter 2
I: Estimated distance to reflector:
I: - Round-Trip Timing method: 2.866765 meters (derived from 8 samples)
I: - Phase-Based Ranging method: 1.402294 meters (derived from 149 samples)
I: Subevent result callback 3
I: Ranging data ready 3
I: Ranging data get completed for ranging counter 3
I: Estimated distance to reflector:
I: - Round-Trip Timing method: 2.922976 meters (derived from 10 samples)
I: - Phase-Based Ranging method: 1.530307 meters (derived from 147 samples)
I: Subevent result callback 4
I: Ranging data ready 4
I: Ranging data get completed for ranging counter 4
I: Estimated distance to reflector:
I: - Round-Trip Timing method: 3.692898 meters (derived from 11 samples)
I: - Phase-Based Ranging method: 1.755326 meters (derived from 146 samples)

So as you can see, the trip time is done using about 10 samples, and the phase-based method is done around 150 times.

I am not completely sure, but I suspect it is done in the procedure_params in main(). It doesn't set the numbers directly, but some intervals and lengths and things like that.

Please note that this sample is not a demo on how good the distance measurement is (because it is pretty bad at this point in time). It is a demonstration of using the technology and fetching the required data. In the future, there will be some more work on the maths used to calculate the actual distance, so that it will be more similar to the performance of the nrf_dm (distance measurement) sample, which uses a proprietary (but similar) technique.

Best regards,

Edvin