NGSIM美国高速公路车辆轨迹数据集

NGSIM美国高速公路车辆轨迹数据集是美国联邦公路局搜集的美国高速公路行车数据，数据集主要包括US101、I-80等道路上的所有车辆在一个时间段内的车辆行驶状况，共包含6个txt文件。该数据集基于楼顶安装的固定摄像头所采集的交通视频，数据本身不涉密。数据使用时，可通过计算机视觉方法从视频中提取出车辆轨迹，并通过Matlab或Python将轨迹数据切割成固定时长的轨迹段，进而用于微观轨迹预测及宏观交通流预测。

The NGSIM US Highway Vehicle Trajectory Dataset is traffic driving data collected by the US Federal Highway Administration (FHWA). The dataset mainly covers the driving states of all vehicles on roads such as US101 and I-80 within a specific time period, and contains a total of 6 TXT files. This dataset is derived from traffic videos captured by fixed cameras installed on building rooftops, and the data itself is not classified. When utilizing this dataset, vehicle trajectories can be extracted from the videos via computer vision methods, and the trajectory data can be segmented into fixed-duration trajectory segments using Matlab or Python, which can then be applied to microscopic trajectory prediction and macroscopic traffic flow prediction.

搜集汇总

背景与挑战

背景概述

该数据集是美国联邦公路局搜集的高速公路车辆轨迹数据，涵盖US101、I-80等道路，包含6个txt文件，总数据量1.15GB，基于摄像头视频提取，适用于自动驾驶轨迹预测和交通流分析。数据由东南大学发布，主要用于微观和宏观交通预测研究。