Looking into vehicle trajectory data, lane changes etc
This project is looking into visualising and quantifying vehicle trajectory available from https://data.transportation.gov/Automobiles/Next-Generation-Simulation-NGSIM-Vehicle-Trajector/8ect-6jqj
The Peachtree data has been found to have a error in the "Global_Time" variable so analysis is going to change to a different area until this has been sorted out - see the "data munging" script for how the error was found.
Different Approaches possible:
-
Looking at individual behaviours (the finest grained x/y of individual vehicles) Dangerous drivers Driver "types" Efficiency of driving How close vehicles are to one another
-
Group behaviours (aggregating from individual sets) Congestion Are driver "types" using certain O/D pairs/routes?
-
Predictions How would lane/road changes effect behaviour?