Create a solution to match trajectories from different tracking sources

[henrymartin1]

I often encounter the problem that I process tracking data from the same tracking study participant that stems from different tracking data sources.
A first example was the SBB Green Class project where study participants had a GPS tracking app installed on their phone as well as in their car. Here the data from the car could be used for a plausibility check of the user labeled transport modes.

I think that we should create a similaritiy module similar to the one [proposed by Sven] (#13) but more general. E.g., not limited to trajectory similarity metrics.
Their should be different types of similarty measurements available (depending of the level of information that is available). Examples are:

• Temporal similarity
  • Temporal overlap
  • max difference between start/end times
• spatial similarity
  • trajectory distance metrics
• semantic similariy

We should distinguish cases where the user ID for both tracking data sources is known and cases where it is not known.