Human Demonstrations for Multi-Modal Task

该数据集包含了人类在执行一项多模态任务时的演示，任务内容是从碗中舀起汤并将其运送到目标位置，同时记录了在执行任务过程中各种传感器的状态。此外，该数据集还捕捉了在多种干扰下任务的动态特性，并详细包含了关于模式不变性和可达性失败的信息。这项研究的任务是基于示范在多模态机器人任务中的学习。

This dataset contains human demonstrations for a multimodal robotic task, where the task requires scooping soup from a bowl and transporting it to a target position. It also records the states of various sensors throughout the task execution process. Additionally, the dataset captures the dynamic properties of the task under multiple disturbances, and includes detailed information about pattern invariance and reachability failures. The core objective of this research is to learn multimodal robotic tasks based on these demonstrations.