Correct classification of fractures according to their patterns is critical for developing a treatment plan in orthopaedic surgery. Unfortunately, for proximal humeral fractures (PHF), methods for proper classification have remained a jigsaw puzzle that has not yet been fully solved despite numerous proposed classifications and diagnostic methods. Recently, many studies have suggested that three-dimensional printed models (3DPM) can improve the interobserver agreement on PHF classifications. Moreover, Virtual Reality (VR) has not been properly studied for classification of shoulder injuries. The current study investigates the PHF classification accuracy relative to an expert committee when using either 3DPM or equivalent models displayed in VR among 36 orthopaedic surgery residents from different hospitals. We designed a multicentric randomised controlled trial in which we created two groups: a group exposed to a total of 34 3DPM and another exposed to VR equivalents. Association between classification accuracy and group assignment (VR/3DPM) was assessed using mixed effects logistic regression models. The results showed VR can be considered a non-inferior technology for classifying PHF when compared to 3DPM. Moreover, VR may be preferable when considering possible time and resource savings along with potential uses of VR for presurgical planning in orthopaedics.
Joint pick and place tasks occur in many interpersonal scenarios, such as when two people pick up and pass dishes. Previous studies have demonstrated that low-dimensional models can accurately capture the dynamics of pick and place motor behaviors in a controlled 2D environment. The current study models the dynamics of pick-up and pass decisions within a less restrictive virtual reality mediated 3D joint pick and place task. Findings indicate that reach-normalized distance measures, between participants and objects/targets, could accurately predict pick-up and pass decisions. Findings also reveal that participants took longer to pick-up objects where division of labor boundaries were less obvious and tended to pass in locations maximizing the dyad's efficiency. This study supports the notion that individuals are more likely to engage in interpersonal behavior when a task goal is perceived as difficult or unattainable (i.e., not afforded). Implications of findings for human-artificial agent interactions are discussed.
During concept design of new vehicles, work places, and other complex artifacts, it is critical to assess positioning of instruments and regulators from the perspective of the end user. One common way to do these kinds of assessments during early product development is by the use of Digital Human Modelling (DHM). DHM tools are able to produce detailed simulations, including vision. Many of these tools comprise evaluations of direct vision and some tools are also able to assess other perceptual features. However, to our knowledge, all DHM tools available today require manual selection of manikin viewpoint. This can be both cumbersome and difficult, and requires that the DHM user possesses detailed knowledge about visual behavior of the workers in the task being modelled. In the present study, we take the first steps towards an automatic selection of viewpoint through a computational model of eye-hand coordination. We here report descriptive statistics on visual behavior in a pick-and-place task executed in virtual reality. During reaching actions, results reveal a very high degree of eye-gaze towards the target object. Participants look at the target object at least once during basically every trial, even during a repetitive action. The object remains focused during large proportions of the reaching action, even when participants are forced to move in order to reach the object. These results are in line with previous research on eye-hand coordination and suggest that DHM tools should, by default, set the viewpoint to match the manikin’s grasping location.
Recent advances in large scale language models have significantly changed the landscape of automatic dialogue systems and chatbots. We believe that these models also have a great potential for changing the way we interact with robots. Here, we present the first integration of the OpenAI GPT-3 language model for the Aldebaran Pepper and Nao robots. The present work transforms the text-based API of GPT-3 into an open verbal dialogue with the robots. The system will be presented live during the HRI2023 conference and the source code of this integration is shared with the hope that it will serve the community in designing and evaluating new dialogue systems for robots.
Recently, exoskeletons have been gaining popularity in many industries, primarily for supporting manual assembly tasks. Due to the relative novelty of exoskeleton technologies, knowledge about the consequences of using these devices at workstations is still developing. Digital human modelling (DHM) and ergonomic evaluation tools may be of particular use in this context. However, there are no standard integrations of DHM and ergonomic assessment tools for assessing exoskeletons. This paper proposes a general method for evaluating the ergonomic effects of introducing an exoskeleton in a production context using DHM simulation tools combined with a modified existing ergonomic assessment framework. More specifically, we propose adapting the Assembly Specific Force Atlas tool to evaluate exoskeletons by increasing the risk level threshold proportionally to the amount of torque that the exoskeleton reduces in the glenohumeral joint. We illustrate this adaptation in a DHM tool. We believe the proposed methodology and the corresponding workflow can be helpful for decision-makers and stakeholders when considering implementing exoskeletons in a production environment.
Remote design reviews are often carried out using video conferencing apps and are limited by the lack of immersive interaction, which is believed to be addressable by using extended reality (XR). It is argued that giving design review participants control over their viewpoint through XR might enhance the design review process. This study investigates whether enhancing camera control can improve collaborative problem-solving without XR. We propose that the ability to create one’s own cognitive map of a space through self-navigation is the basis for improvement, not XR technology specifically.
The experimental setup involves a collaborative puzzle-solving task with two distinct conditions: one with fixed camera perspectives and another allowing personal camera control. Teams of three engage in a task requiring the assembly of a 3D puzzle, where two of them have half of the solution and work to guide a third individual in a puzzle assembly task.
We aim to measure outcomes in terms of completion time, the number of errors, and user satisfaction. Preliminary results indicate a complex interaction between camera control and collaborative dynamics. I intend to discuss our methodology, share initial observations, and explore the implications of these findings.
Aims: Recent world events have resulted in companies using remote meeting tools more often in design processes. The shift to remote meeting tools has had a notable impact on collaborative design activities, such as design reviews (DRs). When DRs depend on computer-aided design (CAD) software, the lack of direct support for CAD functionalities in videoconferencing applications introduces novel communication challenges, i.e. friction. This study investigates friction encountered in real world remote DRs when using a combination of standard CAD and videoconferencing applications. The objective was to understand the main sources of friction when carrying out DRs using a combination of CAD and videoconferencing applications.
Methods: At a single Swedish automobile manufacturer, 15 DRs of a fixture component were passively observed. These observations were subjected to a qualitative thematic analysis to identify categories and sources of friction during these DRs. The DRs were carried out using a combination of CATIA CAD software and Microsoft Teams for videoconferencing.
Results: The analysis of the 15 remote DRs identified four recurring friction categories: requesting specific viewpoints, indicating specific elements, expressing design change ideas, and evaluating ergonomics. Each category highlights specific challenges that were observed during the DRs and emerged due to constraints imposed by existing methods and technologies for remote meetings.
Conclusion: This study provides a framework for understanding the current sources of friction in remote DRs using videoconferencing tools. These insights can support the future development of DR software tools, guiding the integration of features that address these friction points. Additionally, the results serve as a guideline for organizations to implement methods that reduce friction in remote DRs and improve DR quality and efficacy.
Often new digital tools are introduced alongside existing tools and workflows to augment and fill gaps in current processes. Virtual and augmented reality (XR) tools are currently being deployed in this way within design processes, allowing for interactive visualization in virtual environments including the use of DHM tools. Currently, the focus is on how to implement XR as a stand-alone tool for single-user scenarios. However, in collaborative design contexts, screen-based and XR tools can be used together to leverage the benefits of each technology maximizing the potential of multi-user design processes. XR allows for an immersive exploration of designed objects in 3D space, while screen-based tools allow for easier notetaking and integration of additional non-3D software and meeting tools. Ensuring that these technologies are integrated in a mutually beneficial manner requires a framework for determining the best combination of technologies and interfaces for diverse design teams. This paper presents a framework for performing collaborative design reviews in a digital environment that can be accessed using both XR and 2D screen devices simultaneously. It enables asymmetric collaboration to provide each design team member with the technology that best fits their workflow and requirements.
Recently, the concept of Industry 5.0 has been introduced to complement, among other things, Industry 4.0’s focus on efficiency and productivity with a focus on humans in digital design and production processes. The inclusion of human interaction with digital realities, extended reality (XR) technologies, such as augmented reality (AR) and virtual reality (VR), can play an essential role in Industry 5.0. While rapid advances in XR technologies are solidifying and finding their place in the product and production development process, terminology and classification scheme remain under-determined. As a result, there have been numerous classifications of XR technologies from different perspectives, but little widespread agreement. They have been classified by their level of immersion or how well they meet a specific purpose (such as training). In addition to that, the classifications are usually made for one particular field (e.g. marketing, healthcare, engineering, architecture, among others). Therefore, to set the basis for future research, it is essential to identify and outline the dimensions that intervene in product and production design in regards to XR facilitated collaboration. With the ideas proposed in this paper, we want to identify basic concepts that classify a collaborative XR system by analyzing how users interact with the environment and other users. Our motivation is that collaborative design involves not only the physical dimension but also a social dimension. Defining when an XR system contributes to increasing social and/or physical presence could clarify and simplify its categorization.
Tightly coordinated grip force adaptations in response to changing load forces have been reported as continuous, stable, and proportional to the load force changes. Considering the existence of inherent sensorimotor feedback delays, current accounts of grip forceâload force coupling invoke explicit predictive mechanisms in the form of internal models for feedforward control to account for anticipatory grip force modulations. However, recent findings suggest that the stability and regularity of grip forceâload force coupling is less persistent than previously thought. Thus, the objective of the current study was to comprehensively quantify the time-varying characteristics of grip forceâload force coupling. Investigations into the coupling’s dynamics during continuous 30 s bouts of load force oscillation revealed intermittent phases of coordination, as well as phases that varied in stability, rather than a persistent and continuously stable pattern of coordination. These findings have important implications for accounts of grip forceâload force coupling and of anticipation in motor control, more broadly.
The paper reports an investigation conducted during the DHM2020 Symposium regarding current trends in research and application of DHM in academia, software development, and industry. The results show that virtual reality (VR), augmented reality (AR), and digital twin are major current trends. Furthermore, results show that human diversity is considered in DHM using established methods. Results also show a shift from the assessment of static postures to assessment of sequences of actions, combined with a focus mainly on human well-being and only partly on system performance. Motion capture and motion algorithms are alternative technologies introduced to facilitate and improve DHM simulations. Results from the DHM simulations are mainly presented through pictures or animations.
Recent developments in commercial virtual reality (VR) hardware with embedded eye-tracking create tremendous opportunities for human subjects researchers. Accessible eye-tracking in VR opens new opportunities for highly controlled experimental setups in which participants can engage novel 3D digital environments. However, because VR embedded eye-tracking differs from the majority of historical eye-tracking research, in both providing for relatively unconstrained movement and stimulus presentation distances, there is a need for greater discussion around methods for implementation and validation of VR based eye-tracking tools. The aim of this paper is to provide a practical introduction to the challenges of, and methods for, 3D gaze-tracking in VR with a focus on best practices for results validation and reporting. Specifically, first, we identify and define challenges and methods for collecting and analyzing 3D eye-tracking data in VR. Then, we introduce a validation pilot study with a focus on factors related to 3D gaze tracking. The pilot study provides both a reference data point for a common commercial hardware/software platform (HTC Vive Pro Eye) and illustrates the proposed methods. One outcome of this study was the observation that accuracy and precision of collected data may depend on stimulus distance, which has consequences for studies where stimuli is presented on varying distances. We also conclude that vergence is a potentially problematic basis for estimating gaze depth in VR and should be used with caution as the field move towards a more established method for 3D eye-tracking.
Humans commonly engage in tasks that require or are made more efficient by coordinating with other humans. In this paper we introduce a task dynamics approach for modeling multi-agent interaction and decision making in a pick and place task where an agent must move an object from one location to another and decide whether to act alone or with a partner. Our aims were to identify and model (1) the affordance related dynamics that define an actor’s choice to move an object alone or to pass it to their co-actor and (2) the trajectory dynamics of an actor’s hand movements when moving to grasp, relocate, or pass the object. Using a virtual reality pick and place task, we demonstrate that both the decision to pass or not pass an object and the movement trajectories of the participants can be characterized in terms of behavioral dynamics model. Simulations suggest that the proposed behavioral dynamics model exhibits features observed in human participants including hysteresis in decision making, non-straight trajectories, and non-constant velocity profiles. The proposed model highlights how the same low-dimensional behavioral dynamics can operate to constrain multiple (and often nested) levels of human activity and suggests that knowledge of what, when, where and how to move or act during pick and place behavior may be defined by these low dimensional task dynamics and, thus, can emerge spontaneously and in real-time with little a priori planning.
Many common tasks require or are made more efficient by coordinating with others. In this paper we investigate the coordination dynamics of a joint action pick-and-place task in order to identify the behavioral dynamics that underlie the emergence of human coordination. More precisely, we introduce a task dynamics approach for modeling multi-agent interaction in a continuous pick-and-place task where two agents must decide to work together or alone to move an object from one location to another. Our aims in the current paper are to identify and model (1) the relevant affordance dynamics that underlie the selection of the different action modes required by the task and (2) the trajectory dynamics of each actor’s hand movements when moving to grasp, relocate, or pass the object. We demonstrate that the emergence of successful coordination can be characterized in terms of behavioral dynamics models which may have applications for artificial agent design.
Interactive or collaborative pick-and-place tasks occur during all kinds of daily activities, for example, when two or more individuals pass plates, glasses, and utensils back and forth between each other when setting a dinner table or loading a dishwasher together. In the near future, participation in these collaborative pick-and-place tasks could also include robotic assistants. However, for human-machine and human-robot interactions, interactive pick-and-place tasks present a unique set of challenges. A key challenge is that high-level task-representational algorithms and preplanned action or motor programs quickly become intractable, even for simple interaction scenarios. Here we address this challenge by introducing a bioinspired behavioral dynamic model of free-flowing cooperative pick-and-place behaviors based on low-dimensional dynamical movement primitives and nonlinear action selection functions. Further, we demonstrate that this model can be successfully implemented as an artificial agent control architecture to produce effective and robust human-like behavior during human-agent interactions. Participants were unable to explicitly detect whether they were working with an artificial (model controlled) agent or another human-coactor, further illustrating the potential effectiveness of the proposed modeling approach for developing systems of robust real/embodied human-robot interaction more generally.
The human musculoskeletal system’s inherent redundancies allow for infinite potential configurations for any given task. While sometimes seen as a problem for cognitive control systems, motor redundancy also fosters adaptability, learning, and resilience, making it essential for effective motor functioning (Latash, 2012). While many features of human motion and pose production have been identified, it remains unclear how cognitive systems quickly identify and enact motions given the scale of challenges introduced by motor redundancy. This study introduces an inverse kinematics solver, the Forward and Backward Reaching Inverse Kinematics solver (FABRIK) (Aristidou et al., 2016; Lamb et al., 2022). FABRIK uses a novel and lightweight approach to overcoming degree of freedom redundancy in multi-joint systems and may provide insights into human motor control. Initial validations of FABRIK for predicting human motion and pose data, demonstrate strong alignment with recorded data and are comparable to more computationally intensive state-of-the-art methods. We consider the implications of this relatively simple inverse kinematics solver for understanding how cognitive systems might deal with the challenges of motion planning in real time.
Recent research on human-machine interaction (HMI) across a range of fields, including both cognitive science and theatre, has stressed the need to re-frame such interactions as relational and based in shared experience (Gaggioli et al., 2021; Sciutti et al., 2018). In this case, the machine, whether software or hardware based, is characterized as an interaction partner instead of a tool. Reconceiving HMI as involving reciprocity and shared experiences moves away from transactional or one-sided models of interaction and requires exploring what can be meant by reciprocation and shared experience with a non-human partner. In particular, the concept of shared experience in HMI has been relatively under-explored due to both the typical framing of trust in HMI research and technological limitations of HMI systems. Refocusing the design of HMI systems on the ethos of shared experience can be supported by interdisciplinary research with theater.
Posture/motion prediction is the basis of the human motion simulations that make up the core of many digital human modeling (DHM) tools and methods. With the goal of producing realistic postures and motions, a common element of posture/motion prediction methods involves applying some set of constraints to biomechanical models of humans on the positions and orientations of specified body parts. While many formulations of biomechanical constraints may produce valid predictions, they must overcome the challenges posed by the highly redundant nature of human biomechanical systems. DHM researchers and developers typically focus on optimization formulations to facilitate the identification and selection of valid solutions. While these approaches produce optimal behavior according to some, e.g., ergonomic, optimization criteria, these solutions require considerable computational power and appear vastly different from how humans produce motion. In this paper, we take a different approach and consider the Forward and Backward Reaching Inverse Kinematics (FABRIK) solver developed in the context of computer graphics for rigged character animation. This approach identifies postures quickly and efficiently, often requiring a fraction of the computation time involved in optimization-based methods. Critically, the FABRIK solver identifies posture predictions based on a lightweight heuristic approach. Specifically, the solver works in joint position space and identifies solutions according to a minimal joint displacement principle. We apply the FABRIK solver to a seven-degree of freedom human arm model during a reaching task from an initial to an end target location, fixing the shoulder position and providing the end effector (index fingertip) position and orientation from each frame of the motion capture data. In this preliminary study, predicted postures are compared to experimental data from a single human subject. Overall the predicted postures were very near the recorded data, with an average RMSE of 1.67°. Although more validation is necessary, we believe that the FABRIK solver has great potential for producing realistic human posture/motion in real-time, with applications in the area of DHM.
Multiagent activity is commonplace in everyday life and can improve the behavioral efficiency of task performance and learning. Thus, augmenting social contexts with the use of interactive virtual and robotic agents is of great interest across health, sport, and industry domains. However, the effectiveness of humanâmachine interaction (HMI) to effectively train humans for future social encounters depends on the ability of artificial agents to respond to human coactors in a natural, human-like manner. One way to achieve effective HMI is by developing dynamical models utilizing dynamical motor primitives (DMPs) of human multiagent coordination that not only capture the behavioral dynamics of successful human performance but also, provide a tractable control architecture for computerized agents. Previous research has demonstrated how DMPs can successfully capture human-like dynamics of simple nonsocial, single-actor movements. However, it is unclear whether DMPs can be used to model more complex multiagent task scenarios. This study tested this human-centered approach to HMI using a complex dyadic shepherding task, in which pairs of coacting agents had to work together to corral and contain small herds of virtual sheep. Humanâhuman and humanâartificial agent dyads were tested across two different task contexts. The results revealed (i) that the performance of humanâhuman dyads was equivalent to those composed of a human and the artificial agent and (ii) that, using a âTuring-likeâ methodology, most participants in the HMI condition were unaware that they were working alongside an artificial agent, further validating the isomorphism of human and artificial agent behavior.
Extended Reality (XR) is a powerful tool to create new and engaging learning environments. As the technology matures, it opens possibilities for professional training programs where information can be situated in the environment, giving detailed guidance to the user. While this detailed guidance has the potential to help users quickly complete complex tasks, recent research from cognitive psychology indicates that active memory retrieval, often referred to as the testing effect, plays a key role in learning. Specifically, increased support during learning is associated with reduced memory retrieval, with negative effects on long-term retention. While these findings are robust for tasks such as word-pair and image learning, less is known about the impact of the testing effect on motor-skill learning of the type often exercised in XR. In this paper, we present the results of a literature review looking at the state of research on the testing effect related to motor-skill learning and retention. While few articles present findings on the testing effect in motor learning, existing results indicate that the impact of the testing effect on motor learning is similar to non-motor learning; however, more research is necessary in order to draw any strong conclusions.
This paper presents an interview study aiming to understand the state of the art of how ergonomics designers work in the vehicle development process within the Swedish automotive industry. Ten ergonomic designers from seven different companies participated in the interview study. Results report the ergonomics designers' objectives, workflow, tools, challenges, and ideal work performance tool. We identify four main gaps and research directions that can enhance the current challenges: human behavior predictions, simulation tool usability, ergonomics evaluations, and integration between systems.
DHM tools have been widely used to analyze and improve vehicle occupant packaging and interior design in the automotive industry. However, these tools still present some limitations for this application. Accurately characterizing seated posture is crucial for ergonomic and safety evaluations. Current human posture and motion predictions in DHM tools are not accurate enough for the precise nature of vehicle interior design, typically requiring manual adjustments from DHM users to get more accurate driving and passenger simulations. Manual adjustment processes can be time-consuming, tedious, and subjective, easily causing non-repeatable simulation results. These limitations create the need to validate the simulation results with real-world studies, which increases the cost and time in the vehicle development process. Working with multiple Swedish automotive companies, we have begun to identify and specify the limitations of DHM tools relating to driver and passenger posture predictions given predefined vehicle geometry points/coordinates and specific human body parts relationships. Two general issues frame the core limitations. First, human kinematic models used in DHM tools are based on biomechanics models that do not provide definitions of these models in relation to vehicle geometries. Second, vehicle designers follow standards and regulations to obtain key human reference points in seated occupant locations. However, these reference points can fail to capture the range of human variability. This paper describes the relationship between a seated reference point and a biomechanical hip joint for driving simulations. The lack of standardized connection between occupant packaging guidelines and the biomechanical knowledge of humans creates a limitation for ergonomics designers and DHM users. We assess previous studies addressing hip joint estimation from different fields to establish the key aspects that might affect the relationship between standard vehicle geometry points and the hip joint. Then we suggest a procedure for standardizing points in human models within DHM tools. A better understanding of this problem may contribute to achieving closer to reality driving posture simulations and facilitating communication of ergonomics requirements to the design team within the product development process.
Occupant packaging design is usually done using computer-aided design (CAD) and digital human modelling (DHM) tools. These tools help engineers and designers explore and identify vehicle cabin configurations that meet accommodation targets. However, studies indicate that current working methods are complicated and iterative, leading to time-consuming design procedures and reduced investigations of the solution space, in turn meaning that successful design solutions may not be discovered. This paper investigates potential advantages and challenges in using an automated simulation-based multi-objective optimization (SBMOO) method combined with a DHM tool to improve the occupant packaging design process. Specifically, the paper studies how SBMOO using a genetic algorithm can address challenges introduced by human anthropometric and postural variability in occupant packaging design. The investigation focuses on a fabricated design scenario involving the spatial location of the seat and steering wheel, as well as seat angle, taking into account ergonomics objectives and constraints for various end-users. The study indicates that the SBMOO-based method can improve effectiveness and aid designers in considering human variability in the occupant packaging design process.
Interpersonal or multiagent coordination is a common part of everyday human activity. Identifying the dynamic processes that shape and constrain the complex, time-evolving patterns of multiagent behavioral coordination often requires the development of dynamical models to test hypotheses and motivate future research questions. Here we review a task dynamic framework for modeling multiagent behavior and illustrate the application of this framework using two examples. With an emphasis on synergistic self-organization, we demonstrate how the behavioral coordination that characterizes many social activities emerges naturally from the physical, informational, and biomechanical constraints and couplings that exist between two or more environmentally embedded and mutually responsive individuals.
The Negative Attitude toward Robots Scale (NARS) is one of the most common questionnaires used in the studies of human-robot interaction (HRI). It was established in 2004, and has since then been used in several domains to measure attitudes, both as main results and as a potential confounding factor. To better understand this important tool of HRI research, we reviewed the HRI literature with a specific focus on practice and reporting related to NARS. We found that the use of NARS is being increasingly reported, and that there is a large variation in how NARS is applied. The reporting is, however, often not done in sufficient detail, meaning that NARS results are often difficult to interpret, and comparing between studies or performing meta-analyses are even more difficult. After providing an overview of the current state of NARS in HRI, we conclude with reflections and recommendations on the practices and reporting of NARS.
The aim of this extended abstract is to discuss how speech and voice in robots could impact user expectations, and how we, within the human-robot interaction (HRI) research community, ought to handle human-like speech both in research and in the development of robots. Human-like speech refers to both emotions that are expressed through speech and the synthetic voice profile by the robot. The latter is especially important as artificial human-like speech is becoming indistinguishable from actual human speech. Together, these characteristics may cause certain expectations of what the robot is and what it is capable of which may impact both the immediate interactions between a user and robot, as well as a user's future interactions with robots. While there are many ethical considerations around robot designs, we focus specifically on the ethical implications of speech design choices as these choices affect user expectations. We believe this particular dimension is of importance because it not only effects the user immediately, but also the field of HRI, both as a field of research and design. The stance on deception may vary across the different domains that robots are used within; for example, there is a wider acknowledgment of deception in scientific research compared to commercial use of robots. Some of this variation may turn on technical definitions of deception for specific areas or cases. In this paper, we will take on a more general understanding of deception as an attempt to distort or withhold facts with the aim to mislead.
Virtual Reality (VR) could be used to develop more representative Digital Human Modeling (DHM) simulations of work tasks for future Operators 4.0. Although VR allows users to experience the manikin as rather realistic in itself, there are still several aspects that need to be considered when shifting from tasks performed in the real world into a virtual one, adding cognitive and user experience (UX) aspects. Currently, there is limited research of UX in VR. The overall aim was to gain deeper insights into how users’ experiences can ultimately help us to improve how VR can aid in DHM. A pilot study examined how users perceived and experienced actions performed by a humanoid hand (manikin) in VR. Users’ perceived presence indicates how well they are immersed in the virtual environment, and Proactive eye gaze (PEG) was used to measure the realism of the virtual hand. The obtained findings indicate some potentially surprising outcomes and some tentative explanations for these are discussed. The lessons learned from this pilot will be used as input to a future larger study that continues to highlight how UX aspects can be useful in a DHM context.
The human–robot interaction (HRI) field goes beyond the mere technical aspects of developing robots, often investigating how humans perceive robots. Human perceptions and behavior are determined, in part, by expectations. Given the impact of expectations on behavior, it is important to understand what expectations individuals bring into HRI settings and how those expectations may affect their interactions with the robot over time. For many people, social robots are not a common part of their experiences, thus any expectations they have of social robots are likely shaped by other sources. As a result, individual expectations coming into HRI settings may be highly variable. Although there has been some recent interest in expectations within the field, there is an overall lack of empirical investigation into its impacts on HRI, especially in-person robot interactions. To this end, a within-subject in-person study () was performed where participants were instructed to engage in open conversation with the social robot Pepper during two 2.5 min sessions. The robot was equipped with a custom dialogue system based on the GPT-3 large language model, allowing autonomous responses to verbal input. Participants’ affective changes towards the robot were assessed using three questionnaires, NARS, RAS, commonly used in HRI studies, and Closeness, based on the IOS scale. In addition to the three standard questionnaires, a custom question was administered to capture participants’ views on robot capabilities. All measures were collected three times, before the interaction with the robot, after the first interaction with the robot, and after the second interaction with the robot. Results revealed that participants to large degrees stayed with the expectations they had coming into the study, and in contrast to our hypothesis, none of the measured scales moved towards a common mean. Moreover, previous experience with robots was revealed to be a major factor of how participants experienced the robot in the study. These results could be interpreted as implying that expectations of robots are to large degrees decided before interactions with the robot, and that these expectations do not necessarily change as a result of the interaction. Results reveal a strong connection to how expectations are studied in social psychology and human-human interaction, underpinning its relevance for HRI research.