UT-TUMTraf-I

This synthetic data replicates subset R2 sequence 03 of the original TUMTraf-I dataset, containing labeled intersection lidar data. It is designed to mirror the physical layout of the target location, Garching bei München intersection in Germany. The sensors are also modeled after the sensor specifications of the original real dataset.

The synthetic data has an average 46,122.08 points per frame in an 80 x 80 x 10 meters^3 region, compared to original data containing 44,312.94 in the same area. It compares similar to real dataset in terms of scene complexity that is roughly 7 objects per frame. The synthetic dataset also has almost identical mean box size equal to 17.09 meters^3 that is almost identical to real mean box, equal to 18.60 meters^3.