CARMA Dataset

该数据集旨在评估人类-机器人交互中的情境接地情况，特别关注具有详细情境注释的多参与者场景。该数据集包含了多种互动场景，如分类水果、倒液体和交接物品等，并提供了单参与者和多参与者（包括机器人）的配置。该数据集涵盖了8个不同设置下的24段记录，其任务重点在于人类-机器人交互中的情境接地问题。This dataset is designed to evaluate situated grounding in human-robot interaction (HRI), with a particular focus on multi-participant scenarios featuring detailed situational annotations. It encompasses a diverse array of interactive scenarios, such as fruit sorting, liquid pouring, and item handover, and provides both single-participant and multi-participant setups including robotic participants. The dataset comprises 24 recorded sessions across 8 distinct setups, with its task focus centered on the problem of situated grounding in human-robot interaction.

搜集汇总

背景与挑战

背景概述

CARMA Dataset是一个专注于人类-机器人交互中情境接地评估的数据集，特别针对多参与者场景，提供详细的情境注释。它包含分类水果、倒液体和交接物品等多种互动场景，涵盖单参与者和多参与者（包括机器人）配置，共8个设置下的24段记录，用于研究交互中的情境理解问题。