Modeling
At last, I’m able to run simulations using statistics from GPS tracks. It took quite a bit of code and some wrestling with the data, but it is working.
We’re looking at the effect of basing trail decisions (which way to go when you get to an intersection) on your route history. Most models are either the typical trip type (where your route is predetermined as you enter the simulation), or based only on the trail you are entering the intersection from.
The idea is to look at how the effect of choosing the number of decisions you look back has on the output, and the likelihood of the data. Formally, this is Markov chaining.
The new TopoFusion simulator can be run with any length of history–anywhere from the 1 to the maximum (typical trips). So it’s pretty flexible.
I’m excited that I’m once again making progress.
