DWM1001 measuring system accuracy +/-5mm on 3000mm possible?

Hi, i would like to make a system which i can use to measure dimensions from a flat area. My idea is to
setup 4 anchors in my workshop and use 1 tag as measuring point. I have a flat undefined 2D shape on my floor and i place the tag on each wanted point from this shape and save the current location. With these points i would like to create an accurate dimensional drawing/shape on my laptop. The largest dimensions from the shape will be 3000mm. Is this possible to make with an accuracy of somewhere around +/- 5mm?

I have been playing with anchors and tags about 1 year ago and are wondering if such a accuracy is possible these days.

Any thoughts much appreciated!

Possible: Yes.
Easy: No.

UWB tends to be highly repeatable, you may get the wrong location for a point but if you go away and return to it you’ll get the same wrong location. With only 4 anchors the basic accuracy won’t be great, certainly not as good as you want.
But the situation you are describing is a very fixed and defined environment. So assuming the object you are measuring has negligible impact on RF signals (it’s not metal or something similarly reflective) and you are clear of it (the tag is on a pole of known length that keeps it away from the object) then you could calibrate the area.
Measure a grid of points and calculate the errors at each location. You can then interpolate between grid points to create a correction for any point in the room.
This has the added advantage that is also corrects for any errors in your anchor location survey.

One thing to watch for, you would need to ensure the tag antenna orientation is completely consistent, for these accuracies rotating the antenna will be a significant error source.

And any significant changes in the room may require a recalibration. Ideally this would all be done without a person in the way, people are large, unpredictable RF absorbers.

Alternatively you could use a far larger number of anchors. With enough anchors the errors average out a lot more.

Hi AndyA,

Thank you very much for your time and thoughts. Good to know it’s possible and I feel it’s worth setting up a test spot for this.
The orientation of the antenna is perhaps the most difficult challenge. In my opinion, I hold the tag ‘measuring pen’ and take a measurement at any desired measuring point and record it. So this gives an unwanted variable, namely my absorption of RF and the direction of the antenna. Any idea if there are other systems that can be used for something like this?

If I had to do that right now I’d cheat a little. I have a robotic total station available. That is an optical survey system that can track and position a reflective prism to within a couple of mm.
It needs line of sight and can have issues if things are moving too quickly or get obscured but it is accurate. We use it to survey the locations of UWB anchors.

But that’s an expensive piece of kit.

If you are outside RTK GPS is in the few hundred $ region these days. While that’s only good for an absolute position accuracy of 2 cm the relative accuracy between two close points is far better, especially with a bit of averaging. But that requires you to be outside with a clear view of the sky.

One other similar things that may work for you: Combine radio and ultrasound. You send out an ultrasound pulse and a radio pulse at the same time. The anchors pick both up. Because sound moves so slowly timings are far less critical. You can assume the radio travels instantly and so the difference in arrival time gives you distance to the tag. Strong air currents will have an impact and ultrasound is short range so very much small areas indoor only. You also need to calibrate for temperature (it changes the speed of sound) but you can easily measure that impact by measuring between anchors.

Finally some of the VR headset and controller tracking systems can be used to give accurate 3d tracking in a confined space. People use these a lot for lower end camera tracing systems.