V2X-Sim
收藏资源简介:
V2X-Sim是由纽约大学创建的综合性模拟多代理感知数据集,专为V2X辅助自动驾驶设计。该数据集提供多代理传感器记录,包括路边单元(RSU)和多辆车辆的协作感知,以及多模态传感器流,支持多种感知任务。数据集创建过程中使用了SUMO和CARLA模拟器,以生成真实的交通流和同步的传感器流。V2X-Sim旨在促进自动驾驶中的协作感知研究,解决长距离或遮挡区域的感知问题,适用于检测、跟踪和语义分割等任务。
V2X-Sim is a comprehensive simulated multi-agent perception dataset created by New York University, specifically designed for V2X-assisted autonomous driving. This dataset provides multi-agent sensor recordings, including collaborative perception from roadside units (RSUs) and multiple vehicles, as well as multimodal sensor streams, supporting various perception tasks. SUMO and CARLA simulators were utilized during the dataset's creation to generate realistic traffic flows and synchronized sensor streams. V2X-Sim aims to facilitate collaborative perception research in autonomous driving, addressing perception challenges in long-distance or occluded areas, and is applicable to tasks such as detection, tracking, and semantic segmentation.




