Thanks for all the postings and answers so far. I have a few questions:
1. Do all the segments in the same track contain the same targets?
2. What is the difference between geolocation_id and sensor_id in the training metadata? They seem to be identical; will that change in the test set?
3. In the explanation image for the metadata you point out that each location can have more than one sensor, but when looking at geolocation_id and sensor_id in the training metadata they seem to be identical. Did you mean that each geolocation_type can have more than a single sensor? Or can the test set have more than a single sensor for each geolocation_id?
4. For clarification - are "experimental" recordings actual recordings in a lab environment, or are they some sort of simulation?
Thanks in advance,
2) It is possible that in one geolocation_id will be more than one sensor, each has its ID. In some geolocations, there is only one radar (sensor).
3) You can see in the auxiliary data that one geolocation_id can have more than one sensor. The test sets may or may not have geolocation_ids with more than a single radar.
4) The experimental recordings are actual recordings of humans in a "controlled natural environment".
"Controlled natural environment" means that the movement of the targets in the area is not random (like in a "natural" and a real situation) but scheduled, planned and managed.
** However, the geolocations of the "Experiment set" are natural like all other geolocations in the training and test set. **
(MAFAT Challenge Team)