Improving Bluetooth Channel Sounding Measurements with nRF54L15 DK and SDK 2.9.0 for Indoor Positioning

[apes25]

Hello everyone,

Our team is developing an indoor positioning solution and is exploring Bluetooth channel sounding as a core technology. We are currently using the nRF54L15 DK and have been working with the channel sounding examples provided in Nordic Semiconductor SDK version 2.9.0.

During our initial testing, we've observed significant noise in our measurements, resulting in an error of approximately 4 meters. This is impacting the accuracy of our results, and we are seeking guidance on the following:

• Noise Reduction and Configuration: Are there specific configuration parameters or settings within the sample code that can be adjusted to minimize noise and improve measurement accuracy?
• Scalability with Multiple Tags and Anchors: To achieve broader coverage and more robust positioning, we are investigating the feasibility of using multiple reflectors acting as tags and multiple initiators serving as anchors. How would implementing this multi-tag, multi-anchor architecture affect system performance and complexity? and wich parameter shoud we adjust on the example to achive it?

We understand that channel sounding is currently an experimental feature in SDK 2.9.0, but we are keen to investigate its potential for our application. Our objective is to achieve reliable indoor positioning with reasonable accuracy for real-time device tracking.

Any examples, documentation, or shared experiences would be greatly appreciated.

Thank you in advance for your assistance!

Best regards,

[Kenneth]

Hello,

I am not sure what algorithms you are using here, your own or those in the samples. The algorithm in the samples are not very accurate and are mostly to test data flow of between radio and application. So switching algorithm should likely help.

It's possible to setup several anchors and reflectors, this is the same as for normal BLE connections. Be aware that all tracking occurs in connected mode, so this will limit the number. The number of connections will take up time, so you need to balance the number of connections and intervals, this will be application dependent so I am not able to give any best case numbers here on the amount of devices etc.

Hope that helps,
Kenneth

[Vinay Bhagure]

What kind of algorithms or switching algorithms are we talking here about? Can you please elaborate. Thankyou!

-Vinay

[Kenneth]

Hello again,

The Bluetooth Channel Sounding specification covers what is needed to perform all the lower layers of the ranging part. This includes the tone exhanges for PBR and package exchange for the RTT. This results in datasets that is then forwarded up to the application layer.

This is not a range, this is the raw data from the radio activity. These datapoints have to then be fed to an algorithm for getting the range out. There are many ways to do this with different performance and requirements to the system.

The algorithms in the nRF Connect SDK are designed to validate the data flow through the system and are not expected to yield accurate ranging, that was not in scope for these (short development time). 

We are working on having more accurate algorithms in the SDK but for timelines you need to talk with your RSM. You will also have the option to make your own algorithms or use algorithms from 3rd party. These can then be tailored to the environment you are planning to operate in.

Hope that helps,
Kenneth

[Allen_]

When can the more precise ranging algorithm be added to ncs, and can it be provided in the next version? Thank you.

[Kenneth]

You need to contact a Regional Sales Manager (RSM), if you don't know who that is, please use the Sales related questions:
https://www.nordicsemi.com/About-us/Contact-Us 

Kenneth