Collaborative Exploration in Human-Robot Teams; What's in Their Corpora of Dialog, Video, LIDAR Messages?

摘要

This paper briefly sketches new work-in-progress (ⅰ) developing task-based scenarios where human-robot teams collabora-tively explore real-world environments in which the robot is immersed but the humans are not, (ⅱ) extracting and constructing "multi-modal interval corpora" from dialog, video, and LIDAR messages that were recorded in ROS bagfiles during task sessions, and (ⅲ) testing automated methods to identify, track, and align co-referent content both within and across modalities in these interval corpora. The pre-pilot study and its corpora provide a unique, empirical starting point for our longer-term research objective: characterizing the balance of explicitly shared and tacitly assumed information exchanged during effective teamwork.