A Model of Embodied Communications with Gestures between Human and Robots
Generate an AI Snapshot to get a quick, structured summary of this paper.
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
A model of embodied communications that can give an explanation for the mechanism of communications described above and provide evidence for the validity of the model through psychological experiments is proposed.
Abstract
A Model of Embodied Communications with Gestures between Humans and Robots Tetsuo Ono ([email protected]) ATR Media Integration & Communications Research Laboratories 2-2 Hikaridai, Seikacho, Sorakugun, Kyoto, 619-0288 Japan Michita Imai ([email protected]) ATR Media Integration & Communications Research Laboratories Hiroshi Ishiguro ([email protected]) Faculty of Systems Engineering, Wakayama University Abstract In this paper, we propose a model of embodied commu- nications focusing on body movements. Moreover, we explore the validity of the model through psychologi- cal experiments on human-human and human-robot com- munications involving giving/receiving route directions. The proposed model emphasizes that, in order to achieve smooth communications, it is important for a relationship to emerge from a mutually entrained gesture and for a joint viewpoint to be obtained by this relationship. The experiments investigated the correlations between body movements and utterance understanding in order to con- firm the importance of the two points described above. We use robots so that we can control parameters in ex- periments and discuss the issues related to the interaction between humans and artifacts. Results supported the va- lidity of the proposed model: in the case of human-human communications, subjects could communicate smoothly when the relationship emerged from the mutually en- trained gesture and the joint viewpoint was obtained; in the case of human-robot communications, subjects could understand the robot’s utterances under the same condi- tions but not when the robot’s gestures were restricted. Introduction Why do people use gestures when communicating? A common scene on a street involving giving/receiving route directions is some person and a stranger making gestures together as if dancing synchronously and rhyth- mically (see Figure 2). These gestures appear not only when the person describes turns at visible locations, but also at invisible ones. Moreover, it has been shown that people are unable to achieve smooth communications if they are restricted from using spontaneous gestures (Ono et al., 2001). Consequently, gestures play an important role in human-human communications. In this paper, however, we do not discuss emblem ges- tures such as the OK sign; these gestures have arbitrar- ily defined meanings and figures in social conventions. The target of our research is mutually entrained gestures, where a speaker and a hearer spontaneously and syn- chronously move their bodies according to the entrain- ment resulting from mutual actions and utterances. We focus on such gestures because smooth communications between humans can be expected when these gestures are used, as illustrated by the above example involving giving/receiving route directions. In order to investigate the mechanism of the mutu- ally entrained gestures described above, we conduct ex- periments on human-robot communications as well as human-human communications. The reason why we use a robot is that we can unrestrictedly design experiments by using a programmable robot’s gestures. Moreover, an investigation of human-robot communications can con- tribute to research on the methodology of robot design and the interaction between humans and artifacts. The purpose of this paper is to propose a model of embodied communications that can give an explanation for the mechanism of communications described above and, moreover, provide evidence for the validity of the model through psychological experiments. The main characteristic of our model is to focus on the relation- ship emerging from a mutually entrained gesture and the joint viewpoint obtained by the relationship. The exper- iments concretely investigate the correlations between body movements and utterance understanding in human- human and human-robot communications involving giv- ing/receiving route directions. In such a task, it is hard for a person and a stranger to communicate with each other if they do not share the same viewpoint. Here, in order to obtain a joint viewpoint, both sides need to con- struct a relationship emerging from mutually entrained body movements. We investigate the process of commu- nications in the experiments by using our implemented robot. Embodied Communications Previous Research on Gestures Research on gestures conducted to investigate the mech- anism of communications emerged around 1980. In this field, McNeil was the first to carry out cutting- edge research. He pointed out that gestures are syn- chronized with speech in communications, and thus both are closely connected in the cognitive system (McNeil, 1987). McNeil’s research provided findings leading to the development of research on the functions of gestures. However, previous research has mainly analyzed the speaker’s gestures in communications. In other words, many researchers have analytically investigated the cor- relations between speech and the speaker’s body move- ments. Consequently, their aim has been to explain an internal mechanism of an individual speaker. However,
