Open MARS Dataset
收藏资源简介:
Open MARS Dataset是由纽约大学与自动驾驶公司May Mobility合作创建的大型多模态数据集,专注于多代理、多遍历和多模式的自动驾驶研究。该数据集包含来自多次遍历的传感器数据,如GPS轨迹、LiDAR点云和环绕视图RGB图像,覆盖了多种天气和光照条件下的同一地理位置。数据收集过程中,使用了装备有多种传感器的自动驾驶车辆,在特定地理区域内进行多次遍历,以捕捉不同条件下的环境信息。Open MARS Dataset的应用领域包括自动驾驶车辆的感知、预测和规划能力的提升,以及多代理协作感知和学习等新兴研究挑战。
The Open MARS Dataset is a large-scale multimodal dataset co-developed by New York University and autonomous driving firm May Mobility, focusing on autonomous driving research covering multi-agent scenarios, multi-pass traversals and multimodal domains. This dataset includes sensor data from multiple traversals, such as GPS trajectories, LiDAR point clouds and surround-view RGB images, covering the same geographical location under diverse weather and lighting conditions. During the data collection process, autonomous vehicles equipped with various sensors were used to conduct multiple traversals within a specific geographical area, capturing environmental information under different conditions. The application fields of the Open MARS Dataset include improving the perception, prediction and planning capabilities of autonomous vehicles, as well as tackling emerging research challenges like multi-agent collaborative perception and learning.

- 1Multiagent Multitraversal Multimodal Self-Driving: Open MARS Dataset纽约大学 · 2024年



