On the Importance of Environments in Human-Robot Coordination

TL;DR

This paper emphasizes the critical role of environment design in human-robot collaboration, introducing a procedural generation framework that creates diverse, solvable environments to study their impact on coordination behaviors.

Contribution

It presents a novel procedural environment generation method that ensures diversity, solvability, and stylistic similarity, enabling systematic study of environmental effects on human-robot coordination.

Findings

Environments lead to qualitatively different behaviors.

Significant differences in collaboration metrics observed.

Environment influences coordination even with same robot algorithms.

Abstract

When studying robots collaborating with humans, much of the focus has been on robot policies that coordinate fluently with human teammates in collaborative tasks. However, less emphasis has been placed on the effect of the environment on coordination behaviors. To thoroughly explore environments that result in diverse behaviors, we propose a framework for procedural generation of environments that are (1) stylistically similar to human-authored environments, (2) guaranteed to be solvable by the human-robot team, and (3) diverse with respect to coordination measures. We analyze the procedurally generated environments in the Overcooked benchmark domain via simulation and an online user study. Results show that the environments result in qualitatively different emerging behaviors and statistically significant differences in collaborative fluency metrics, even when the robot runs the same planning algorithm.