More than a Million Ways to Be Pushed

本数据集名为‘More than a Million Ways to Be Pushed’，由麻省理工学院的计算机科学与人工智能实验室创建。数据集包含超过一百万条时间戳标记的推力实验数据，涵盖了推力平面、推力对象的姿态以及交互力。数据集通过工业机器人自动捕获，精确控制推力的位置、速度和加速度，覆盖了多种推力交互维度，如表面材料、对象形状、接触位置、推力方向、推力速度和推力加速度。该数据集旨在通过实验研究推力，为机器人操作中的摩擦模型提供实验基础，解决推力预测中的不准确性问题。

This dataset, titled "More than a Million Ways to Be Pushed", was constructed by the Computer Science and Artificial Intelligence Laboratory (CSAIL) at the Massachusetts Institute of Technology (MIT). It comprises over one million timestamped push experiment datasets, encompassing push planes, poses of the pushed objects, and interaction forces. Automatically collected via industrial robots, the dataset features precise control over the position, velocity, and acceleration of the applied pushes. It covers diverse push interaction dimensions, including surface materials, object geometries, contact positions, push directions, push velocities, and push accelerations. This dataset is designed to support experimental research on robotic pushing, provide an experimental basis for friction models in robotic manipulation, and mitigate inaccuracy issues in push prediction.