channel sounding distance estimation

Dear Support Team,

I am currently working with two nRF54L15 DK boards placed approximately 50 cm apart and stationary. I am using the channel_sounding_ras_ example with SDK v2.9.1 to evaluate channel sounding precision.

However, I am encountering significant variations in the measurements, even though the boards remain static. Additionally, I am unable to achieve a precision better than half a meter.

To improve accuracy, I have adjusted several parameters, including:

abstracts from main.c (initiator)

line 437 and 438

		.rtt_type = BT_CONN_LE_CS_RTT_TYPE_128_BIT_RANDOM,
		.cs_sync_phy = BT_CONN_LE_CS_SYNC_2M_PHY, 


line 476 and 477

		.snr_control_initiator = BT_LE_CS_INITIATOR_SNR_CONTROL_30dB,
		.snr_control_reflector = BT_LE_CS_REFLECTOR_SNR_CONTROL_30dB,

Despite these modifications, the precision remains poor, and I occasionally observe large fluctuations in the results.

Could you provide guidance on how to improve the accuracy and stability of the channel sounding measurements? Are there additional parameters I should adjust, or known limitations with this setup?

Thank you in advance for your support.

Best regards,
Maxime

Parents
  • Hi

    I'm sorry, but the Channel sounding is still very much in experimental mode yet, and the ranging algorithms are very basic, so very accurate ranging can't be expected at this point. We will have better versions coming, but in the end the algorithm for ranging is up to you to improve as it will be outside of what we support in the SDK.

    Best regards,

    Simon

  • I used the RAS example in SDK 3.0.1 and my results are:
    ifft: 1.234041, phase_slope: 1.854832, rtt: 8.811189
    (real distance is 0.8 m, indoor)

    As users, can we only try to improve the result of each of these three methods by changing different parameters and testing? Is that what you mean? Or will there also be some improvements in the SDK libraries, so that we get better results even with the same parameters?

    Can we expect some improvements in the next SDK version? For example, will there be a way to combine these three results into one?

    Can we expect any improvement from Nordic in the future?

    Thank you, Jan.

Reply
  • I used the RAS example in SDK 3.0.1 and my results are:
    ifft: 1.234041, phase_slope: 1.854832, rtt: 8.811189
    (real distance is 0.8 m, indoor)

    As users, can we only try to improve the result of each of these three methods by changing different parameters and testing? Is that what you mean? Or will there also be some improvements in the SDK libraries, so that we get better results even with the same parameters?

    Can we expect some improvements in the next SDK version? For example, will there be a way to combine these three results into one?

    Can we expect any improvement from Nordic in the future?

    Thank you, Jan.

Children
No Data
Related