Precision

[arthurwolf]

I've been over the website for a few minutes, and can't find an answer to my question, so I thought I'd ask here.
What kind of resolution does this project provide ? Are we talking dozen of meters ? Meters ? Centimeters ? I get it's variable, but a general idea would tell me if this is a project I want to look into or not.

Cheers :)

[bleckers]

Hi Arthur,

The standard itself doesn't define accuracy, as it's designed to work with different hardware as well as a plethora of different distance measurement techniques (for example it can work with 802.11mc's fine time measurement when that's released). The whole purpose of this was to offer a dataless method of transferring the access point location to a client, so it can figure out its position without data connectivity (and the reduced latency that comes from not having to process the position on an external server).

However, I have tested it in a controlled space with 8 Wi-Fi nodes/access points and got up to 0.5 meter accuracy with a smartphone. Though this was a fixed environment and was tweaked to the hardware I was using, so using different hardware would change this.

The beauty of this system is that it's backwards compatible and easy to setup on existing access points, so it's something you can test quickly to see if it suits. However with a mildly dense constellation of access points (say one per room/area), you will never really get anything better than knowing you were in that room. So accuracy really depends on the setup and use case of the implementer.

Having said this, I am developing an experimental and hobbyist system for better accuracy. I have just put the finishing touches on the firmware for the new hardware and am getting comparable accuracy to ultra wide band systems for a much lower cost, and plan to release this soon - https://hackaday.io/project/9242-subpos-ranger. Just note that this isn't really designed for large scale commercial setups and it will require special hardware on the client. But it is suitable for small scale hobbyist and robotic setups.

Hopefully this answers your question :)

Regards,

Blair

[arthurwolf]

Thanks a lot, it does answer my question :)
Maybe this has it's place in the FAQ ( that's where I looked first ) ?

On Thu, Jun 9, 2016 at 3:40 AM, Blair Wyatt notifications@github.com
wrote:

Hi Arthur,

The standard itself doesn't define accuracy, as it's designed to work with
different hardware as well as a plethora of different distance measurement
techniques (for example it can work with 802.11mc's fine time measurement
when that's released). The whole purpose of this was to offer a dataless
method of transferring access point location to a client, so it can figure
out its position without data connectivity (and the reduced latency that
comes from not having to process the position on an external server).

However, I have tested it in a controlled space with 8 Wi-Fi nodes/access
points and got up to 0.5 meter accuracy with a smartphone. Though this was
a fixed environment and was tweaked to the hardware I was using, so using
different hardware would change this.

The beauty of this system is that it's backwards compatible and easy to
setup on existing access points, so it's something you can test quickly to
see if it suits. However with a mildly dense constellation of access points
(say one per room/area), you will never really get anything better than
knowing you were in that room. So accuracy really depends on the setup and
use case of the implementer.

Having said this, I am developing an experimental and hobbyist system for
better accuracy. I have just put the finishing touches on the firmware for
the new hardware and am getting comparable accuracy to ultra wide band
systems for a much lower cost, and plan to release this soon -
https://hackaday.io/project/9242-subpos-ranger. Just note that this isn't
really designed for large scale commercial setups and it will require
special hardware on the client. But it is suitable for small scale hobbyist
and robotic setups.

Hopefully this answers your question :)

Regards,

Blair

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
#1 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AAGpFdfpZPMUdaqCQpIwhXlISeLGHc-gks5qJ28cgaJpZM4IxUft
.

Courage et bonne humeur.

[bleckers]

No worries, I will pop it in the FAQ shortly :)