Demonstration Data for Perceptive Locomotion

该数据集包含了从可操控的购物车在不同环境中收集的专家演示数据，旨在通过模仿学习来训练高级导航控制器。此外，该数据集还包括了在不同难度级别下收集的轨迹，这些难度级别包括简单、中等、困难，并且涵盖了各种动态和静态条件。具体规模上，该数据集包含了250条现实世界的轨迹，以及每个难度级别80条演示轨迹。该数据集的任务是研究在动态环境中的感知性运动。

This dataset contains expert demonstration data collected from a maneuverable shopping cart across various environments, aiming to train advanced navigation controllers via imitation learning. Furthermore, the dataset also includes trajectories collected at three difficulty levels: easy, medium and hard, which cover various dynamic and static conditions. In terms of scale, the dataset comprises 250 real-world trajectories, plus 80 demonstration trajectories for each difficulty level. The core task of this dataset is to investigate perceptual motion in dynamic environments.