NatSGD
收藏资源简介:
NatSGD是一个多模态人机交互数据集,由马里兰大学帕克分校创建。该数据集包含1143条通过语音和手势发出的自然人类指令,同步记录了机器人的行为演示。数据集内容涵盖日常厨房任务,如食物准备、烹饪和清洁,旨在训练机器人理解复杂任务。创建过程中采用了Wizard of Oz实验设计,确保交互的自然性。NatSGD的应用领域包括多模态感知、视觉识别、模仿学习等,旨在解决人机交互中的任务理解问题。
NatSGD is a multimodal human-robot interaction dataset created by the University of Maryland, College Park. This dataset contains 1,143 natural human instructions issued via speech and gestures, with synchronously recorded robot behavior demonstrations. The dataset covers daily kitchen tasks including food preparation, cooking and cleaning, aiming to train robots to comprehend complex tasks. The Wizard of Oz experimental design was adopted during its creation to ensure the naturalness of the interaction. The application fields of NatSGD include multimodal perception, visual recognition, imitation learning and other related areas, aiming to address task understanding challenges in human-robot interaction.

- 1NatSGD: A Dataset with Speech, Gestures, and Demonstrations for Robot Learning in Natural Human-Robot Interaction马里兰大学帕克分校 · 2024年



