Extending Commands Embedded in Actions for Human-Robot Cooperative Tasks

Abstract

In this paper, we propose a novel method to control a robot without robot manipulation. Users do not need to precisely manipulate the robot and to learn the manipulation method. The proposed method can send commands to a robot by using a human action sequence that achieves their own task. In order to enable the robot to achieve tasks, we introduce a keepbased interaction in which a human keeps an action in the sequence for a certain period. The advantages of our method are efficiency improved by not requiring additional human actions, and functionality to enable a robot to perform further actions. We consider that the efficiency is supported by users' physical workloads and cognitive loads. Users' physical workloads would not be increased because the proposed method does not require additional human actions. However, the effect of the proposed method on users' cognitive loads is unknown. We applied the method to a desktop sweeping task by a human and a small mobile robot, and conducted an experiment with participants to measure users' cognitive loads in a cooperative sweeping task. As a result, we found that the proposed method had a lower cognitive load than a typical conventional method.