Vehicle Energy Dataset (VED)
收藏资源简介:
Vehicle Energy Dataset (VED)是由密歇根大学创建的一个大规模数据集,包含从2017年11月至2018年11月期间,在美国密歇根州安娜堡收集的383辆个人汽车的燃油和能量数据。该数据集捕捉了车辆的GPS轨迹以及燃油、能量、速度和辅助电源使用的时间序列数据。数据集中的车辆类型多样,包括264辆汽油车、92辆混合动力车和27辆插电式混合动力/电动车。VED数据集总里程约374,000英里,涵盖了从高速公路到交通密集的市中心区域等各种驾驶条件和季节。数据集创建过程中,研究团队通过安装在车辆上的OBD-II记录器收集数据,并对个人身份信息进行了去标识化处理,以保护参与者隐私。VED数据集的应用领域广泛,包括车辆能源消耗建模、驾驶员行为建模、机器学习和深度学习、交通模拟器的校准、最佳路线选择模型、人类驾驶员行为预测以及自动驾驶汽车的决策制定等。
Vehicle Energy Dataset (VED) was created by the University of Michigan. It is a large-scale dataset containing fuel and energy data of 383 personal passenger vehicles collected in Ann Arbor, Michigan, USA, from November 2017 to November 2018. This dataset captures the GPS trajectories of the vehicles, as well as time-series data on fuel/energy consumption, vehicle speed and auxiliary power utilization. The dataset includes diverse vehicle types: 264 gasoline-powered vehicles, 92 hybrid vehicles, and 27 plug-in hybrid/electric vehicles. The total cumulative mileage of the VED dataset is approximately 374,000 miles, covering various driving conditions and seasons, ranging from highways to traffic-congested downtown areas. During the dataset development process, the research team collected data via OBD-II loggers installed on the participating vehicles, and performed de-identification of personal identifiable information to protect the privacy of the participants. The VED dataset has a wide range of application areas, including vehicle energy consumption modeling, driver behavior modeling, machine learning and deep learning, traffic simulator calibration, optimal route selection models, human driver behavior prediction, and autonomous vehicle decision-making, among others.




