Haptic Human–Robot Collaboration: Comparison of Robot Partner Implementations in Terms of Human-Likeness and Task Performance

Abstract

In the past, working spaces of humans and robots were strictly separated, but recent developments have sought to bring robots into closer interaction with humans. In this context, physical human–robot interaction represents a major challenge, as it is based on continuous bilateral information and energy exchanges which result in a mutual adaptation of the partners. To address the challenge of designing robot collaboration partners, making them as human-like as possible is an approach often adopted. In order to compare different implementations with each other, their degree of human-likeness on a continuous scale is required. So far, the human-likeness of haptic interaction partners has only been studied in the form of binary choices. In this paper, we first introduce methods that allow measuring the human-likeness of haptic interaction partners on a continuous scale. In doing so, two subjective rating methods are proposed and correlated with a task performance measure. To demonstrate the applicability and validity of the proposed measures, they are applied to a joint kinesthetic manipulation task and used to compare two different implementations of a haptic interaction partner: a feedforward model based on force replay, and a feedback model. This experiment demonstrates the use of the proposed measures in building a continuous human-likeness scale and the interpretation of the scale values achieved for formulating guidelines for future robot implementations.