nuPlan
收藏资源简介:
nuPlan数据集是世界上首个真实世界的自动驾驶数据集和基准,由Motional机构创建。该数据集包含1282小时的多样化驾驶场景,涵盖拉斯维加斯、波士顿、匹兹堡和新加坡四个城市,并提供高质量的自动标注物体轨迹和交通灯数据。数据集旨在评估基于机器学习的规划器在处理各种驾驶情况时的能力,确保安全和高效决策。此外,nuPlan还提供了一个模拟和评估框架,允许规划器的动作在闭环中进行模拟,以考虑与其他交通参与者的交互。
The nuPlan dataset is the world's first real-world autonomous driving dataset and benchmark, created by Motional. It contains 1,282 hours of diverse driving scenarios across four cities: Las Vegas, Boston, Pittsburgh, and Singapore, and provides high-quality automatically annotated object trajectories and traffic light data. The dataset is designed to evaluate the capabilities of machine learning-based planners when handling various driving situations, ensuring safe and efficient decision-making. In addition, nuPlan also provides a simulation and evaluation framework that allows closed-loop simulation of planner actions to account for interactions with other traffic participants.




