arduino-dw1000 icon indicating copy to clipboard operation
arduino-dw1000 copied to clipboard

Dw1000 wrong distance and calibration

Open fkaraokur opened this issue 7 years ago • 5 comments

How do I calibrate the dw1000 module? Will we do calibration for anchor and tag? Or are we going to make arrangements in the library? I measure 274 cm with my anchor. But the actual distance is 196.5.

fkaraokur avatar Aug 19 '17 17:08 fkaraokur

Same issue here. For now I just subtract a constant to compensate. I am wondering and searching what's the cause of this problem and how to fix also. Thanks for bringing up this question. Will update if I have anything new.

kerr4huo avatar Aug 19 '17 22:08 kerr4huo

I think this is a chip fiasco. Because every time I switch the device on and off, the metrics change. So distance travel is not static but dynamic. That's why I can not develop a formula. I expect help from the master.

fkaraokur avatar Aug 20 '17 13:08 fkaraokur

I have the same issue. The error varies between 0.5 to 0.75 meters. Do you have any idea how i can correct it please?

werty37 avatar Sep 11 '17 18:09 werty37

Do you have any idea how i can correct it please?

Hi @werty37 In my project I used a practical approach. I've measured the error for various distances (2-20m). Each measurement is represented by red plus symbol.

The error varies between 0.5 to 0.75 meters.

You're right here. I have the same issue. In general, I would say that DW1000 chip is inaccurate but quite precise (the results are repeatable). So it's not that bad ;)

Using this points I generate a function. I called it 'correction function'. I used 'polyfit' function in MATLAB to get this blue line on the graph.

Now we can finally make an use of the correction function.

  1. DW1000 measures 7,65m
  2. The measured value is the input of correction function. Correction function calculates it. The output is about 0,79m.
  3. The DW1000 chip substracts 7,65-0,79 and displays 6,86m. image

I just subtract a constant to compensate. @kerr4huo

My method is an extension of your idea. It does NOT remove the measuring error but help to reduce it.

Piotr

prideofisland avatar Sep 23 '17 14:09 prideofisland

Hi @prideofisland

That is a good idea. I will try to follow proper anchor positioning and then try your method. As of now i am trying to make the Ranging Anchor code on an ESP8266 since i will do the localization on the PC. If i am able to port it successfully, i will create a wiki/howto. Lets see.

Thanks S

werty37 avatar Sep 26 '17 20:09 werty37