I am working on testing anchors to find a drop off that would be good to measure near patient/far from patient.
I did the tests on the old and new anchors ranging from .5 – 3 meters away (in steps of .5), from 0 – 180 degrees (in steps of 20), both vertically and horizontally oriented.
(BRUB = .5 m, KI3U = 1 m, KICX = 1.5 m, BSK1 = 2 m, BS7K = 2.5 m, BSIX = 3 m)
Here are the Cumulative Density Functions from the data:
The new anchors, horizontally oriented gives little distinction for us to use for near patient/far from patient.
Using new anchors, vertically gives us better distinction in the middle range.
Using the old anchors, horizontally oriented doesn’t give us the distinction between near patient/far from patient that we’re looking for.
From the CDF graphs, we can see that the old anchors, vertically oriented gives us the best distinction between near patient/far from patient, but this could be improved. We would want better distinction between 1 and 1.5 meters. Also, we want a higher RSSI peak within boxplots per angle, so that there would be an area around the bed (oval-shaped) that would count as near the bed and everything else would be far from bed.