TerraSentia Robot Data

该数据集是通过TerraSentia机器人在玉米和大豆田中收集的，目的是为了评估所提出异常检测模型的性能。该数据集包含了机器人不同操作模式的标注类别（正常、行进碰撞、不可通行障碍、可通行障碍），并用于训练自编码变分模型（SVAE），以便进行故障识别。这些数据是在多次机器人运行中收集的，其任务是对机器人导航中的异常进行检测。

This dataset was collected in corn and soybean fields by the TerraSentia robot, with the objective of evaluating the performance of the proposed anomaly detection model. It includes annotated categories corresponding to different operation modes of the robot, namely normal, traveling collision, impassable obstacle, and passable obstacle, and is used to train the Stochastic Variational Autoencoder (SVAE) for fault identification. All data was gathered across multiple robot runs, where the task was to detect anomalies during robot navigation.