added relative observations
added testing for relative observations

TODO - multiple robots (modify datastructures), math
stays the same
Tested on both simulated and real-world data, works in both cases

Small tweaks in parameter passing in simulated dm

(only had to modify generation of groundtruths)
changed initial starting states of robots in dm
added some basic path plotting code to test_simulation

Added a data-generator class to
generate data for simulated_dataset-manager