It’s raining today so I cannot test outside So instead I wanted to learn more about GDOP. I wrote a python script to calculate GDOP, HDOP, and VDOP from my four anchors and tag position. I followed the math in wiki and here. After fiddling around a bit it seems to work. GDOP and VDOP get worse/larger as I move my tag further away, and improves/smaller when I increase the separation of my tags along the z-axis (shifting one pair higher than the other pair).

So that’s interesting. But I don’t really know what these DOPs mean. I was wondering if there was a way to convert to estimated error in meters? Or if I’m just supposed to use them as unitless metrics for “that makes it worse”.

I believe that in theory if you scale all the elements in the Q matrix by the square of the standard deviation of the range measurement errors (assuming the ranges are on average correct) then the results will be in meters. At least that’s my understanding, my understanding of the maths isn’t great.

Having said that I’ve only ever used it as a relative measure to tell me that configuration A is better than B. The theory never seemed to match reality that well and I ended up with a far better way of assessing accuracy in real time.

Since I’m looking at large areas with anchors coming in and out of range I have a tool that takes the anchor locations, maximum usable range and expected tag height. It then plots the number of anchors in range and DoP at every point in the site. Calculating the DoP for 12 anchors at every point on a 500x500 grid takes a while, don’t try that in python if you’re in a rush. You very quickly learn to predict exactly what the results will be so the benefit is minimal for frequent users but it gives a nice graphical visualisation for people unfamiliar with the system.

One thing that isn’t intuitive (or wasn’t to me) but makes sense when you think about it is that the VDOP improves significantly if you can move all the anchors significantly above the tag.