SceNDD
收藏资源简介:
SceNDD是一个基于场景的自然驾驶数据集,由交通与自动驾驶系统研究所(TASI)在印第安纳波利斯市中心通过装备车辆收集。数据集包含68个驾驶会话,每个会话由不同驾驶员进行,持续20至40分钟。数据集内容包括自车及非自车的轨迹点、速度、偏航角等信息,以及非自车的进入和退出时间。创建过程中,使用了联合概率数据关联(JPDA)跟踪器来检测道路上的非自车。数据集的应用领域主要集中在开发高效的运动规划和路径跟踪算法,旨在解决自动驾驶车辆在复杂交通环境中的导航问题。
SceNDD is a scenario-based natural driving dataset collected by the Institute of Transportation and Autonomous Driving Systems (TASI) using instrumented vehicles in downtown Indianapolis. This dataset contains 68 driving sessions, each conducted by a distinct driver, with durations ranging from 20 to 40 minutes. The dataset includes information such as trajectory points, speed, and yaw angle of both the ego vehicle and non-ego vehicles, as well as the entry and exit times of non-ego vehicles. During its creation, a Joint Probabilistic Data Association (JPDA) tracker was employed to detect non-ego vehicles on the road. The primary application scenarios of this dataset focus on developing efficient motion planning and path tracking algorithms, aiming to address the navigation challenges of autonomous vehicles in complex traffic environments.




