I'm a student and I am working on a GPS-like system. I have implemented a few algorithms (Newton-Raphson, Bancroft, ...) and I would like to compare them.
Basically, I have a constellation of satellites and some receivers, and I want to quantify the capacity of each algorithm to compute (x,y,z,clock_bias) for each receiver. See the disposition below.
I want to introduce a plot with errors against a criterion based on the disposition of satellites around receivers.
I tried to work with a GDOP (Geometric Dilution of Precision), but the results look weird, sometimes between 0 and 1, sometimes between 5 and 6, but the configuration of satellites has not really changed.
So do you have any idea for a criterion with the hypothesis all parameters (satellites and receivers), bias include, are known?
