Faless/harvest_apples_with_agilex_piper_sim

该数据集是一个用于机器人学研究的模拟数据集，由LeRobot工具创建。数据集聚焦于苹果采摘任务，使用Agilex Piper机器人模型进行模拟。它包含600个完整的情节（episodes），总计1,020,000帧数据，以25帧每秒的速率采集。数据集结构包括观测数据（如机器人关节位置、夹持器力矩，以及来自两个视角的视觉图像：ego视角和front视角，分辨率均为360x640像素）、动作数据（机器人关节和夹持器的位置控制）、时间戳、帧索引、情节索引等。数据以parquet文件格式存储，视频以mp4格式存储，总数据大小约为300MB。数据集旨在支持机器人控制、视觉感知和任务学习的研究，适用于训练和评估机器人算法。

This dataset is a simulation dataset for robotics research, created using the LeRobot tool. It focuses on an apple harvesting task, simulated with the Agilex Piper robot model. The dataset contains 600 complete episodes, totaling 1,020,000 frames collected at 25 frames per second. The structure includes observations (such as robot joint positions, gripper torque, and visual images from two perspectives: ego and front, both with a resolution of 360x640 pixels), actions (position control for robot joints and gripper), timestamps, frame indices, episode indices, and more. Data is stored in parquet format, with videos in mp4 format, and the total data size is approximately 300MB. The dataset is designed to support research in robot control, visual perception, and task learning, suitable for training and evaluating robotic algorithms.