Channel sounding performance results

[ing-barbiz]

Hello,

we have been testing channel sounding with single antenna configuration, however we are not sure of which accuracy to expect. Also, we are not sure of the reliability of the measurement.

We will try also the 1x1 configuration with the Pixel Phone and the 2x2 (board-to-phone and board-to-board) as well with the nRF54 TAG.

However, we ask if you have any reference measurement test data to share. So we can be sure we are correctly reproducing the results.

Thanks

Marco

[Elfving]

Hi,

I might be able to dig some up, though there are some things I want to get out of the way first.

So first of all, CS isn't UWB; the best you can expect is ±0.5m. Differentiating between ±0.5m would be impossible/hard. That is just describing CS from a technical and theoretical stand-point. 

Another point is what our SDK provides is mainly a BLE compliant way of getting the I/Q values needed to estimate the distance. From there you are free to estimate the distance in multiple ways that might have some drawbacks (precision vs computational power etc.). We do provide one algorithm for this in the SDK, but there are multiple others available online, and we do not claim to provide the best one. I would rather urge you to ask your local Nordic sales manager (RSM) for who you can contact in order to get a better algorithm.

If you want the best results from the sample, it is tempting to simply say that you should disregard anything but the IFFT measurements from the sample, as IFFT might be the most developed algorithm we have there, but the truth is rather that all of these metrics work best in different situations. RSSI is okay in short range for instance, and RTT is best for very long ranges (and I wonder if the phase slope might get you a max range, but don't quote me on that). This old webinar (which is meant to describe how our old "Distance Toolbox" worked, but the tech is almost the same) goes a bit further into the details of what separates these measurements. 

So are you comparing distance estimations or I/Q values here? The latter is rather difficult by itself. The NDT webinar I linked to above lists some ideal measurements though, channel sounding is assumed to perform a bit better than that.

Regards,

Elfving

[ing-barbiz]

Dear Elfving,

thanks much for your complete answer.

We are aware of algorithm providers and we are already contacting one.

Also, thanks for the clarification about the IFFT.

We are comparing distance estimates. But, as in any experiment, first of all you want to establish a baseline. Establishing a baseline means repeating another one's results.

So my question is if you can share an experiment record to validate our results.

Thanks,

Marco

[Elfving]

This are some quick ones I just threw together (using for instance several meter long cables instead of a measuring tape), in an office environment with people walking around. 

NCS 3.3 and 54L15DKs.

Let me know if you were interested in a specific distance I didn't test here, or wanted something more precise.

1m

2m

3m

5m

Regards,

Elfving

[ing-barbiz]

Hello Elfving,

also from your results, I understand the measurement is not very reliable.

There is both offset error and noise.

Such figures do not allow to use the results for user interaction purposes.

So I will ask to some company prividing their proprietary algorithms and let's see.

Thanks

Marco