This paper presents the theoretical basis for understanding how transparency utterances in social robots can be systematically employed to regulate trust. Current approaches to trust calibration often lack more detailed insights into how different verbal explanations of actions or intentions shape users' mental models of the robot's capabilities. Drawing on pragmatic principles and mechanisms from human interaction research, we present a model of how transparency utterances contribute to mental models of the robotic interaction partner. We test the model in a user study on the effects of transparency utterances that allow inferences about higher- or lower-level capabilities. The in-person, between-subject experiment with N=47 shows that depending on the design of the transparency utterance, users perceive the competence, benevolence and transparency of the robot differently and infer more or fewer additional capabilities. The results confirm the effect of the proposed design of transparency utterances on (over-)trust. Our model thus offers a theoretical foundation for developing transparency strategies in various contexts of human-robot interaction.

• We present three different studies, starting off from six different existing notation systems and then refining and testing the most promising visualizations and modes of presentation in two further studies. • We combine different methodologies by carrying out think-aloud usability test, thus collecting both qualitative statements and behavioral data on student performance, which we compare with a large-scale reception experiment. • We find that iconic systems are easier to interpret than symbolic ones; • we find that stylized contours may lead to better productions than concrete contours; • we find that combining stress and intonation in the same notation is not a problem in itself and that a two-step presentation, in which the information on the stress pattern of the utterance is presented first, can help the integration of information on stress and intonation; • nevertheless, we find that the notation system we propose, which combines stress information with stylized intonation contours, turns out most robust, leading reliably to good learner productions. In this paper, we present three studies in which we develop and test different representations of intonation contours and prosodic stress with the aim to provide foreign language learners with a notation that they can easily interpret. Study I addresses learners’ productions of target intonation based on six common notation systems. The results reveal how learners themselves make sense of the different representations and show that learners produce fewest errors in iconic representations with stylized contours. Based on these results, Study II examines the role of complexity of information by experimenting with one- versus two-step presentations of the teaching material. While Study I was carried out as a production task by pairs of learners, Study II shows in an online perception test with native speakers that a stylized representation with a sparse representation of stress performs best. Study III explores these findings in another production task. The results confirm that the most suitable visualization technique of those investigated is to present stylized intonation contours with the main points of emphasis added, and that a two-step presentation can be helpful.

This study examines whether the humanoid robot, HuGo, can effectively influence human behavior in routine scenarios, using receipt collection at a university cafeteria as a case study. The robot encourages customers to remember to take their receipts at the self-checkout register, as negligence to do so creates additional workload for cafeteria employees. The research tests whether using verbal and pointing gestures enhances compliance compared to verbal-only or text-based prompts. The testing is conducted as an in-the-wild experiment, and the data is analyzed with both Chi-squared test and Two-Proportion Z-test. Results indicate that introducing speech makes a significant improvement in user compliance, while the influence of hand gestures is inconclusive.

Much work on human-robot interaction has shown that such interactions can profit from implementing human-like behaviors, in line with theoretical approaches that assume that human-like social cues ’trigger’ or ’evoke’ social behaviors towards the respective robotHowever, there is also evidence that people treat interactions with robots in special ways, that they have different expectations and attend to different communicative tasks than in interactions with other humans; especially those features that are geared towards efficiency in interaction seem not to be relevant or even perceived positively in human-robot interaction. In this paper, we investigate the effects of the relative timing of emotional expression while speaking; in a controlled in-person interactive experiment with N=56, participants interacted with a simulated robot that either presented certain emotional behaviors after the respective utterance or timed with the main content units during speech, which had been determined empirically in a prior study of interactions between humans. Results show that even though the ill-timed emotional expressions cause interruptions and problems with respect to turn-taking, participants prefer the robot that plays emotional behaviors after the utterance – thus deprioritizing the efficiency and turn-taking requirements of human interaction. The results thus support a constructive perspective on human-robot interaction, where participants engage in sophisticated sense-making based on the character depicted and their own understanding of the interaction situation.

In this work, we analyze video data and interviews from a public deployment of two trash barrel robots in a large public space to better understand the sensemaking activities people perform when they encounter robots in public spaces. Based on an analysis of 274 human–robot interactions and interviews with N = 65 individuals or groups, we discovered that people were responding not only to the robots or their behavior, but also to the general idea of deploying robots as trashcans, and the larger social implications of that idea. They wanted to understand details about the deployment because having that knowledge would change how they interact with the robot. Based on our data and analysis, we have provided implications for design that may be topics for future human–robot design researchers who are exploring robots for public space deployment. Furthermore, our work offers a practical example of analyzing field data to make sense of robots in public spaces.

In this position paper, I argue that both HRI academia and robot developers can profit from defining interaction as coordination. Viewing human-robot interactions as instances of coordination provides not only a novel starting point for the evaluation of interactions, but also a practical tool for robot development. Using the example of the development of a service robot application, I show that an approach that puts interaction as coordination first provides an easy access to interaction design and defines clear goals. Furthermore, I argue that ethnomethodological conversation analysis constitutes a rigorous analytical tool to study the extent to which a given robot is able to achieve coordination at various levels, thus providing opportunities for HRI research and evaluation.

Many social robots will have the capacity to interact via speech in the future, and thus they will have to have a voice. However, so far it is unclear how we can create voices that fit their robotic speakers. In this article, we explore how robot voices can be designed to fit the size of the respective robot. We therefore investigate the acoustic correlates of human voices and body size. In Study I, we analyzed 163 speech samples in connection with their speakers’ body size and body height. Our results show that specific acoustic parameters are significantly associated with body height, and to a lesser degree to body weight, but that different features are relevant for female and male voices. In Study II, we tested then for female and male voices to what extent the acoustic features identified can be used to create voices that are reliably associated with the size of robots. The results show that the acoustic features identified provide reliable clues to whether a large or a small robot is speaking.

In this study, we examine whether touching a soft robot while doing different tasks can make participants synchronize their breathing rhythm with the robot. 28 participants interacted with the robot, which either was inflated and deflated, thus simulating breathing, or remained inactive. During the experiment, data were collected through two breathing belts and an EEG device. The findings of the study suggest higher arousal associated with positive emotional valence for participants in the breathing robot condition compared to the inactive robot condition. The participants in the breathing robot condition also breathed more deeply and regularly and blinked fewer times, a finding that suggests lower stress levels in comparison with people who interacted with the inactive robot. The analysis of the data suggests that touching the breathing robot led to some degrees of stress reduction, yet without leading to synchronization with the robot's inhalation rhythm.

In interaction with robots, verbal interaction is important and can have an impact on the perception of the robot. Experiences from a pilot study showed that the KUBO robot was not intuitive to use. To investigate if tailored verbal utterances could make the robot more intuitive to use, a new usability test combined with a Wizard of Oz method, where a facilitator played verbal utterances when KUBO drove over the specific TagTiles, was carried out. The test shows that the tailored verbal utterances helped in the understanding of KUBO and the actions of the robot when passing over the TagTiles which therefore made it more intuitive to use. Even though the verbal utterances helped the users, there were still some TagTiles they did not understand. The test also showed that there was a considerable mismatch between what the participants stated they understood and what was observed they really understood.

This paper compares how much people trust in a robotic system and in a group of people. Our experiment focuses on trust in robot knowledge versus trust in collective human knowledge tested with the help of a quiz game. During the experiments, people are competing with each other and gain or lose points based on their decisions. A joker system was designed that provides evidence on whether people rely rather on the robot or on human collective knowledge. In this setup, perceived intelligence is therefore highly correlated with trustworthiness. In the scope of this joker-picking system, overall results show that the trust in the robotic system is higher than the trust in collective human knowledge.

Humans are an essential part of human-robot interaction (HRI), but what roles do they play in HRI research? Analysis of the role of human subjects in research can serve as an indicator of how the HRI community engages with society. In this paper, we examine humans' roles in the HRI studies published at the ACM HRI conference over the course of 16 years (between 2006–2021). We categorize the studies into three groups. The studies in the first group investigated human nature and studied humans as interchangeable subjects; the studies in the second group addressed humans as users of robots in certain contexts; the third group of studies approached humans as social actors who are closely connected to other actors and thereby generate social dynamics. The contributions of this paper are twofold: First, we reveal the patterns of how humans have been included in HRI studies. Specifically, we find that more than half of the studies limited the role of humans to interchangeable and generalizable actors. Second, we outline three opportunities for the HRI community that arise if human subjects are given more diversified roles in HRI research - opportunities for diversity, social justice, and reflexivity. On this basis, we call for a more socially -engaged research in HRI.

Existing methodologies to describe anthropomorphism in human-robot interaction often rely either on specific one-time responses to robot behavior, such as keeping the robot's secret, or on post hoc measures, such as questionnaires. Currently, there is no method to describe the dynamics of people's behavior over the course of an interaction and in response to robot behavior. In this paper, I propose a method that allows the researcher to trace anthropomorphizing and non-anthropomorphizing responses to robots dynamically moment-by-moment over the course of human-robot interactions. I illustrate this methodology in a case study and find considerable variation between participants, but also considerable intrapersonal variation in the ways the robot is anthropomorphized. That is, people may respond to the robot as if it was another human in one moment and to its machine-like properties in the next. These findings may influence explanatory models of anthropomorphism.

This article is a commentary on Hoffman 8 Zhao’s “A Primer for Conducting Experiments in Human-robot Interaction.” I argue that a too-narrow view of HRI methodology fails to address the dynamic systems properties of interaction. Furthermore, the focus on addressing the so-called “replicability crisis” makes field studies next to impossible, inhibits interdisciplinarity and methodological pluralism, and draws our attention and resources away from the fact that contexts, people, cultures, expectations, and interaction itself may influence how social signals are interpreted. Therefore, in spite of its great benefits, the “Primer” may not be taken as an instruction on “how to carry out research in HRI” in general.

Incremental feedback, i.e. the timely response to human behavior while it is happening, has previously been found to potentially speed up human-robot interactions, but it is unclear how people evaluate incremental robots. In this study, we show that the evaluation of incremental robot response depends on the actual success of the incremental feedback; that is, if the feedback leads to increased efficiency, people evaluate the robot as more competent and more credible. If the robot does not use incremental feedback, no interaction between evaluation and efficiency can be found. Thus, incremental feedback draws people's attention to interaction success.

Field studies where robots are tested in real life settings bring different challenges for researchers, robotics scientists and users. In this paper, we address some of the challenges we encountered when testing two different drink-serving service robots in the wild. We collect challenges from three different experiments. Two experiments were conducted in elderly care facilities, while a third experiment took place in the lobby of a concert hall. We focus on the challenges that researchers face during the preparation phase and the on-set deployment phase when testing robots in the wild. We point to potential difficulties that may arise and present some practical solutions to the issues encountered. Our results suggest that lab studies do not sufficiently prepare the researcher for research 'in the wild'.

In this article, we address to what extent the proverb “the sound makes the music” also applies to human-robot interaction, and whether robots could profit from using speech characteristics similar to those used by charismatic speakers like Steve Jobs. In three empirical studies, we investigate the effects of using Steve Jobs’ and Mark Zuckerberg's speech characteristics during the generation of robot speech on the robot's persuasiveness and its impressionistic evaluation. The three studies address different human-robot interaction situations, which range from online questionnaires to real-time interactions with a large service robot, yet all involve both behavioral measures and users’ assessments. The results clearly show that robots can profit from using charismatic speech.

In this paper, we draw attention to the social functions of emotional display in interaction. A review of HRI papers on emotion suggests that this perspective is rarely taken in the field, but that it is useful to account for the context- and culture-dependency of emotional expression. We show in two case studies that emotional display is expected to occur at very specific places in interaction and rather independently from general emotional states, and that different cultures have different conventions regarding emotional expression. Based on conversation analytic work and the results from our case studies, we present design recommendations which allow the implementation of specific emotional signals for different human-robot interaction situations.

In this paper, we test whether the perception of filled-pause (FP) frequency and public-speaking performance are mediated by the phonetic characteristics of FPs. In particular, total duration, vowel-formant pattern (if present), and nasal segment proportion of FPs were correlated with perceptual data of 29 German listeners who rated excerpts of business presentations given by 68 German-speaking managers. Results show strong inter-speaker differences in how and how often FPs are realized. Moreover, differences in FP duration and nasal proportion are significantly correlated with estimated (i.e. subjective) FP frequency and perceived speaker performance. The shorter and more nasal a speaker's FPs are, the more do listeners underestimate the speaker's actual FP frequency and the higher they rate the speaker's public-speaking performance. The results are discussed in terms of their implications for FP saliency and rhetorical training.

In medical human-robot interactions, trust plays an important role since for patients there may be more at stake than during other kinds of encounters with robots. In the current study, we address issues of trust in the interaction with a prototype of a therapeutic robot, the Universal RoboTrainer, in which the therapist records patient-specific tasks for the patient by means of kinesthetic guidance of the patients arm, which is connected to the robot. We carried out a user study with twelve pairs of participants who collaborate on recording a training program on the robot. We examine a) the degree with which participants identify the situation as uncomfortable or distressing, b) participants' own strategies to mitigate that stress, c) the degree to which the robot is held responsible for the problems occurring and the amount of agency ascribed to it, and d) when usability issues arise, what effect these have on participants' trust. We find signs of distress mostly in contexts with usability issues, as well as many verbal and kinesthetic mitigation strategies intuitively employed by the participants. Recommendations for robots to increase users' trust in kinesthetic interactions include the timely production of verbal cues that continuously confirm that everything is alright as well as increased contingency in the presentation of strategies for recovering from usability issues arising.

In this paper, we address issues of timing between robot behaviors in multimodal human-robot interaction. In particular, we study what effects sequential order and simultaneity of robot arm and body movement and verbal behavior have on the fluency of interactions. In a study with the Care-O-bot, a large service robot, in a medical measurement scenario, we compare the timing of the robot's behaviors in three between-subject conditions. The results show that the relative timing of robot behaviors has significant effects on the number of problems participants encounter, and that the robot's verbal output plays a special role because participants carry their expectations from human verbal interaction into the interactions with robots.

Programming by Demonstration (PbD) is an efficient way for non-experts to teach new skills to a robot. PbD can be carried out in different ways, for instance, by kinesthetic guidance, teleoperation or by using external controls. In this paper, we compare these three ways of controlling a robot in terms of efficiency, effectiveness (success and error rate) and usability. In an industrial assembly scenario, 51 participants carried out peg-in-hole tasks using one of the three control modalities. The results show that kinesthetic guidance produces the best results. In order to test whether the problems during teleoperation are due to the fact that users cannot, like in kinesthetic guidance, switch between control points using traditional teleoperation devices, we designed a new device that allows users to switch between controls for large and small movements. A user study with 15 participants shows that the novel teleoperation device yields almost as good results as kinesthetic guidance.

In this paper, we explore experimentally the possible tradeoff between gaze to the user and gaze to the path in robot approach. While some previous work indicates that gaze towards the user increases perceived safety because the user feels recognized, other work indicates that it is legibility of the robot's actions that put users at ease. If the robot does not drive up to the person in a straight line directly, the robot can either continuously look at the person and thus maintain eye contact, or indicate its path through its gaze behavior, increasing legibility. In an experiment with N=36 participants, we tested the tradeoff between legibility and eye contact. The behavioral results show that users are significantly more at ease with the robot that gazes at them than with the robot that looks where it is going, measured by the number of instances of glances away from the robot. Likewise, the participants rate the robot that looks at them continuously as more intelligent and more cooperative. Thus, participants value mutual gaze higher than legibility.

Dialog between drivers and speech-based vehicle interfaces can be used as an instrument to find out what drivers might be concerned, confused or curious about in driving simulator studies. Eliciting on-going conversation with drivers about topics that go beyond navigation, control of entertainment systems, or other traditional driving related tasks is important to getting drivers to engage with the activity in an open-ended fashion. In a structured improvisational Wizard of Oz study that took place in a highly immersive driving simulator, we engaged participant drivers (N=6) in an autonomous driving course where the vehicle spoke to drivers using computer-generated natural language speech. Using microanalyses of the drivers’ responses to the car’s utter- ances, we identify a set of topics that are expected and treated as appropriate by the participants in our study, as well as a set of topics and conversational strategies that are treated as inappropriate. We also show that it is just these unexpected, inappropriate utterances that eventually increase users’ trust in the system, make them more at ease, and raise the system’s acceptability as a communication partner.

Service robots in public places need to both understand environmental cues and move in ways that people can understand and predict. We developed and tested interactions with a trash barrel robot to better understand the implicit protocols for public interaction. In eight lunch-time sessions spread across two crowded campus dining destinations, we experimented with piloting our robot in Wizard of Oz fashion, initiating and responding to requests for impromptu interactions centered on collecting people's trash. Our studies progressed from open-ended experimentation to testing specific interaction strategies that seemed to evoke clear engagement and responses, both positive and negative. Observations and interviews show that a) people most welcome the robot's presence when they need its services and it actively advertises its intent through movement; b) people create mental models of the trash barrel as having intentions and desires; c) mistakes in navigation are indicators of autonomous control, rather than a remote operator; and d) repeated mistakes and struggling behavior polarized responses as either ignoring or endearing.

In this paper, problems in instructing an industrial robot by means of a control panel are investigated. In order for the robot to learn as much and as fast as possible from demonstration, the demonstration by the teacher needs to be as precise as possible. Usability studies constitute a useful methodology to investigate in which situations users provide the robot with exact trajectories and, if not, why they face difficulties. Results show that movements involving only the lowest joint of the robot arm are straight-forward and very exact. In contrast, fine movements that involve joints of the upper arm cause considerable problems. The analysis shows that users have to employ separate actions instead of focusing on the target and therefore need to consciously plan the action since they cannot match the robot's movements with those of their own embodiment.

In this paper, the causes for singularity of a robot arm in teleoperation for robot learning from demonstration are analyzed. Singularity is the alignment of robot joints, which prevents the configuration of the inverse kinematics. Inspired by users' own hypotheses, we investigated speed and delay as possible causes. The results show that delay causes problems during teleoperation though not in direct control with a control panel because users expect a different, more intuitive control in teleoperation. Speed on the other hand was not found to have an effect on the occurrence of singularity.

While robot learning from demonstration comes with great benefits [5], the intuitive interaction between naïve users and robots also poses challenges. For instance, users need to be prevented from causing damage and to be enabled to recover from errors. We studied the error resolution strategies of 28 lay users performing simple assembly tasks via teleoperation of a robotic arm in order to gain insight into the strategies users take. The two most common problems are too much pressure and singularity. Even though users were provided with instructions on how to undo singularity in an instruction video, they did not always recover successfully. In contrast, too much pressure, if noticed, was resolved mostly correctly by lifting the peg or by letting it drop into the hole rather than inserting it. Finally, users were quite clueless about how to resolve self-collision and over-rotation.

Robots often need to ask humans for help, for instance to complete a human component in a larger task or to recover from an unforeseen error. In this paper, we explore how robots can initiate interactions with people in order to ask for help. We discuss a study in which a robot initiated interaction with a participant by producing either an acoustic signal or a verbal greeting. Thereafter, the robot produced a gesture in order to request help in performing a task. We investigate the effect that social framing by means of a verbal greeting may have on people's attention to the robot, on their recognition of the robot's actions and intention, and on their willingness to help. The results show that social framing, in contrast to other methods for getting a person's continued attention, is effective and increases how friendly the robot appears. However, it has little influence on people's willingness to assist the robot, which rather depends on the activities people are engaged in, and on the readability of the robot's request.

In this paper, we investigate the role of physical embodiment of a robot and its degrees of freedom in HRI. Both factors have been suggested to be relevant in definitions of embodiment, and so far we do not understand their effects on the way people interact with robots very well. Linguistic analyses of verbal interactions with robots differing with respect to physical embodiment and degrees of freedom provide a useful methodology to investigate factors conditioning human-robot interaction. Results show that both physical embodiment and degrees of freedom influence interaction, and that the effect of physical embodiment is located in the interpersonal domain, concerning in how far the robot is perceived as an interaction partner, whereas degrees of freedom influence the way users project the suitability of the robot for the current task.

The current investigation addresses whether the socially guided machine learning paradigm can be extended to a new domain, embodied grammar learning. Experimental results show that naive users indeed reduce the complexity of linguistic utterances in tutoring sessions for a simulated robot, even though their own knowledge of the subject area is only tacit. These findings have implications for the usability of robots as `teachable agents', as well as for automatic language learning from interaction.

In this paper, I investigate interpersonal variation in verbal HRI with respect to the computers-as-social-actors hypothesis. The analysis of a corpus of verbal human-robot interactions shows that only a subgroup of the users treat the robot as a social actor. Thus, taking interpersonal variation into account reveals that not all users transfer social behaviors from human interactions into HRI. This casts doubts on the suggestion that the social responses to computers and robots reported on previously are due to mindlessness. At the same time, participants' understanding of robots as social or non-social actors can be shown to have a considerable influence on their linguistic behavior throughout the dialogs.

This paper addresses a so far neglected area of human-robot interaction by approaching situation awareness from the point of view of naive users. In particular, we present an investigation into which naive models of robots' capabilities users carry into the interaction, how these models influence the interaction, and which means can be taken to guide users into more realistic models and behaviours if necessary. Quantitative and qualitative investigations reveal not only considerable uncertainty about robots' situation awareness, but also significant differences in dealing with this uncertainty. Three different types of users can be distinguished on the basis of the strategies they take. Finally, we describe experiments with two means of shaping the users' models into more realistic accounts of robots' capabilities. The results suggest that verbal robot output is a powerful means for guiding users subtly, unobtrusively and online into an understanding of robots' capabilities that is more realistic and adequate than users' naive models of robots' situation awareness.

The question addressed in this paper is which types of spatial reference human users employ in their interaction with a robot and how a cognitively adequate model of these strategies can be implemented. Experiments in human-robot interaction were carried out which show how human users approach an artificial communication partner which was designed on the basis of empirical findings regarding spatial references among humans. The results are considerable differences in the strategies which speakers employ to achieve spatial reference in human-robot interaction and in natural communication.