mbodiai/ABB_PandG_20250401_174952

这个数据集包含了机器人执行任务时的观察数据、动作数据、状态数据、奖励以及元数据。观察数据由图像和字符串指令组成，动作数据包括机器人的姿态（由x、y、z、roll、pitch和yaw六个浮点数表示）、抓握力度（浮点数）和夹爪力度（整型）。状态数据包括两个摄像头视角的图像和深度信息、世界中的物体以及夹爪的姿态（由一系列浮点数表示）。奖励是一个浮点数，表示执行动作后获得的奖励。元数据包含剧集索引和步骤索引。数据集分为训练集，提供了相应的字节数和示例数。

This dataset includes observation, action, state, reward, and metadata during a robots task execution. Observation data consists of images and string instructions, action data includes the robots pose (represented by six floats: x, y, z, roll, pitch, and yaw), grasp force (a float), and gripper force (an integer). State data includes images and depth information from two camera perspectives, objects in the world, and the grippers pose (represented by a sequence of floats). The reward is a float indicating the reward obtained after performing an action. Metadata contains episode index and step index. The dataset is split into a training set, with provided byte size and number of examples.