CODD (Cooperative Driving Dataset)

合作驾驶数据集是使用 CARLA 生成的合成数据集，其中包含来自多个车辆的激光雷达数据，这些车辆同时在一组不同的驾驶场景中导航。该数据集的创建是为了进一步研究多智能体感知（协作感知），包括协作 3D 对象检测、协作对象跟踪、多智能体 SLAM 和点云注册。为实现这一目标，所有帧都标有真实传感器姿势和 3D 对象边界框。

The Cooperative Driving Dataset is a synthetic dataset generated using CARLA, which contains LiDAR data from multiple vehicles navigating concurrently across a set of diverse driving scenarios. This dataset was developed to advance research on multi-agent perception (collaborative perception), including collaborative 3D object detection, collaborative object tracking, multi-agent SLAM, and point cloud registration. To this end, all frames are annotated with ground-truth sensor poses and 3D object bounding boxes.

搜集汇总

背景与挑战

背景概述

CODD是基于CARLA模拟器生成的多车激光雷达合成数据集，旨在支持协作感知研究，如3D检测与跟踪。所有数据均标注了传感器姿态和3D物体边界框。