Non-Rectangular area

I have an application to locate some items in a non-rectangular area. The area is almost a Cross shape.

Would I need to determine imaginary points outside the cross area to use as reference?

ie. say the bottom left corner (which is outside the cross) could be considered 0,0 and the first anchor inside the cross would be say 3,0 (assuming the actual area starts 3m in from the imaginary corner).
And all other anchors would be relative to that datum?


You could do it that way.
Or you could have 0,0 in the middle and have negative values for some areas.

Everything needs to be relative to the same fixed point but where you put that point is up to you.

Thanks Andy,
I put up a sample structure in our building using 6 Tags and a cross format.
I am using the old 1.04? android code and DWM1001 Kit (supplied with 12 nodes).

Though the max. distance in my building is only 5 or 6m between any anchor, the plotting of the Tag (I am only using 1 Tag) is very unreliable.

Is the Decawave module really affected by non line of sight? This would make it impractical for use in warehouses with lots of racking if it only works properly in open spaces with no obstructions.

Or, could the issue be a bug in the android code?

I set one anchor as initiator and other 5 anchors and normal anchors.


It’s only going to be reliably accurate line of sight.

It’s not magic, if the radio signal is blocked how is it going to make a measurement? Even if the signal can pass through the wall it’s going to travel slower through the obstruction than if does through the air and so your range measurement will be inaccurate.
In non-line of sight situations the performance depends a lot on what the obstruction is made of. Metal shelving racks are mostly empty space and cardboard or plastic aren’t much of an issue so in some warehouse situations it would still work fairly well with the view blocked.
If you replaced the cardboard and plastic with metal or liquids then the quality would drop rapidly since the direct signal will be almost completely blocked and you’ll probably end up measuring reflected signal paths.

For a cross shaped area assuming the walls have some structure to them I would say you need at least 8 tags for good performance, 2 at the ends of each arm. That way you always have line of sight to 4. Since you have 12 I’d also put 2 in the middle area.