1. NGSIM dataset
The full name of NGSIM is Next Generation Simulation. It is a data acquisition project initiated by the US Federal Highway Administration. Trajectory prediction, driver intention recognition, autonomous driving decision planning, etc. All the data are the actual running trajectory data collected on the U.S. Highway National Highway 101. The data file contains 3 txt files, and the three files correspond to three different periods, each of which is 15 minutes.
Official download: https://data.transportation.gov/Automobiles/Next-Generation-Simulation-NGSIM-Vehicle-Trajector/8ect-6jqj
2. HighD dataset
The HighD data set released by the Institute of Automotive Engineering of RWTH Aachen University in Germany is a large-scale natural vehicle trajectory data of German highways . It was collected from six different locations near Cologne, Germany. The locations vary according to the number of lanes and speed limits. The recorded The data includes cars and trucks. The dataset includes 11.5 hours of measurements from six locations and 110,000 vehicles with a total mileage of 45,000 km measured and 5,600 complete lane change records. By using state-of-the-art computer vision algorithms, positioning errors are typically less than ten centimeters. It is suitable for tasks such as driver model parameterization, autonomous driving, and traffic pattern analysis.
The data format of HighD is slightly different from that of the subsequent InD and RounD.