nuPlan
收藏资源简介:
nuPlan数据集是由上海交通大学开发的,用于评估端到端自动驾驶模型的闭环反应性模拟基准。该数据集基于大规模真实世界驾驶数据,将长距离驾驶旅程分割成多个可管理的驾驶场景,每个场景包含高层次的导航信息和传感器数据。nuPlan数据集的主要目的是解决现有模拟器在真实性和反应性方面的不足,特别是在高度交互的驾驶场景中。通过提供高保真度的传感器数据和反应性行为模拟,nuPlan数据集为自动驾驶模型的评估和开发提供了强有力的支持。
The nuPlan dataset, developed by Shanghai Jiao Tong University, is a closed-loop reactive simulation benchmark for evaluating end-to-end autonomous driving models. Built on large-scale real-world driving data, this dataset splits long-distance driving journeys into manageable driving scenarios, each containing high-level navigation information and sensor data. The core objective of the nuPlan dataset is to address the shortcomings of existing simulators in terms of realism and reactivity, particularly in highly interactive driving scenarios. By offering high-fidelity sensor data and reactive behavior simulation, the nuPlan dataset provides robust support for the evaluation and development of autonomous driving models.




