Vehicle trajectory data processed from the Waymo Open Dataset

Related paper: Processing, assessing, and enhancing the Waymo autonomous vehicle open dataset for driving behavior research (submitted to the journal Transportation research part C: emerging technologies)

Version 3 Update
1. Fix the issue of negative speed in the trajectores (note that the proportion of negative speeds is very small, it is 0.245% for speed smaller than -0.1m/s). The paired CF trajectory files (with or without large vehicles) have been updated.

Version 2 Content
1. The data processing codes
2. CF pair trajectories where large vehicles are involved are added

Version 1 Content
This dataset is processed from the Waymo Open Dataset (perception part). This dataset includes the following content:
1. Tabular format object information (Autonomous vehicle and objects detected by Lidar such as vehicles, pedestrians, cyclists)
2. Visualized object trajectory (in the format of top view videos)
3. Tabular format paired Car Following (CF) trajectory in which the CF pairs have been carefully selected for CF behavior research