During the past 15 years, an increasing amount of works have investigated proactive robotic behavior in relation to Human–Robot Interaction (HRI). The works engage with a variety of research topics and technical challenges. In this article, a review of the related literature identified through a structured block search is performed. Variations in the corpus are investigated, and a definition of Proactive HRI is provided. Furthermore, a taxonomy is proposed based on the corpus and exemplified through specific works. Finally, a selection of noteworthy observations is discussed.
1 Introduction
During the past four decades, developments within different fields of robotics have introduced purposefully designed robots into the dynamic environment of humans. Within the past two decades, the emergence of collaborative robots has brought industrial manipulators from fixed enclosed industrial environments into the open proxemics of humans. There is now a multitude of commercial domestic and social robots designed to interact with humans; and, likewise, autonomous collaborative robots, in line with the principles of Industry 4.0 and 5.0, have been applied to assist human workers in industry. As a natural consequence of the interest in introducing robots in applications where non-roboticists face them, the field of Human–Robot Interaction (HRI) has emerged. HRI is concerned with all facets of human–robot interaction, including safety aspects and the design of intuitive interfaces accommodating engagement of inexperienced people with technology as complex as robots.
For some time, the focus has been on designing feasible HRI for when a human initiates and dictates a robot’s actions or the robot reacts to certain input. However, within the past 15 years, the interest in pushing the boundaries and obtaining more intelligent and autonomous robotic behavior has introduced the subfield of Proactive Human–Robot Interaction (Proactive HRI), which has been increasingly researched. The field of Proactive HRI derives from the desire of having more intuitive and natural human–robot interactions inspired by how humans interact with each other. In Proactive HRI, current themes such as human-centered solutions and advancement of socially acceptable intelligent autonomous robots are in focus, making it an important field of the times.
2 Related Works
Proactivity has been mentioned in relation to HRI in a fair amount of research; still it seems there has been no clear definition of what can be considered Proactive HRI. When a new field arises, it is subject to the risk that (mis-)uses of the related terms can result in vague understandings of the field. This can be counteracted by clear definitions based on systematic and intentional studies of the field.
In the past, different reviews have been published in the fields of HRI and, more specifically, Proactive HRI. Goodrich and Schultz [60] provided an insight into HRI in their survey, where they strive to discuss and promote a unified understanding and handling within the field. Furthermore, they discuss key themes and important challenges that shape this field. Other authors have published works focused on providing an overview of a specific area within HRI or HRI within a specific domain (e.g., metrics within HRI [169], task planning [179], social robots [36], physical HRI (pHRI) [38], and more recently proactive robots in HRI [166]). While Reference [166] is concerned with Proactive HRI, it is mainly focused on the delimited topic of robots’ emotional intelligence gained from proactivity based on observing nonverbal human cues. Consequently, the article addresses selected aspects of Proactive HRI but does not address Proactive HRI in full as the wide spectrum field it is. Similarly, Bianco and Ognibene [16] and Brinck and Balkenius [22] provide a review of the use of Theory of Mind in Proactive HRI and mutually adaptive interaction, respectively. The work by Peng et al. [140] that studies proactivity in relation to robots offering Decision Making Support presents a division of proactivity into high, medium, and low proactivity levels. However, the given distinctions are context-specific and not generalizable. These observations and the fact that a clear overview of proactivity within HRI is not yet available in the literature motivate this work with the purpose of creating a unified understanding of the nuances of Proactive HRI.
This work reviews previous research studying proactivity in relation to HRI and, based on this, provides a clear definition of what can be considered Proactive HRI. Furthermore, a taxonomy is presented, which can assist researchers to identify and clarify which branch of Proactive HRI they are engaging with, which in turn will result in a more rigorous structure of future work within this field. Last, the state-of-the-art methods described in the literature within each subcategory of the branches are reviewed. This can be summarized in the following contributions:
(1)
A clear definition of Proactive HRI
(2)
A taxonomy for Proactive HRI.
(3)
A review of the state-of-the-art within Proactive HRI with a focus on common tendencies and variation between works.
(4)
An overview of the distribution of the papers within the proposed taxonomy.
The remainder of this article is structured as follows: Section 3 describes the method and procedure used to obtain the relevant literature for this review. Section 4 analyzes the general understandings and structures that can be observed in the included literature, and in Section 5 a taxonomy based on these is provided. Subsequently, highlights in the identified subcategories within the branches of Proactive HRI with respect to focus are provided, and the state-of-the-art within the subcategories is outlined. Section 6 discusses some general cross-category observations, open research questions, as well as the proposed taxonomy. Finally, Section 8 summarizes and concludes this review.
3 The Corpus of Literature
The relevant literature for mapping out the field of Proactive HRI was identified through a structured approach planned and documented using a search protocol. Important considerations from the protocol will be detailed in this section. A block search was conducted on several databases utilizing search terms identified through an initial “Quick and Dirty” free text search on Google Scholar. Since the field to cover is wide, a two-aspect block search with the following terms was constructed: Aspect 1: Proactiv* (where the asterisk ensures inclusion of form of the word in the search), Aspect 2: Human-Robot Interaction, HRI, Human-Robot Collaboration, Human-Robot Cooperation. The searches are constructed in this way to obtain all combinations of proactivity with relevant broad terms related to interactions.
Fig. 1.
Fig. 1. The amount of hits found on the individual databases and the resulting amount of articles after sorting and duplicates removed.
A walkthrough of all available databases, to our knowledge, and their coverage was conducted resulting in the identification of a total of eight relevant databases and one net-crawler (see Figure 1). The selected databases were chosen to include archives with an area-specific coverage, e.g., IEEE Xplore, ACM Digital Library, as well as an interdisciplinary coverage; Scopus, Web of Science, and so on. Block searches were constructed on each database and, when possible, predetermined inclusion and exclusion criteria were applied. The inclusion criteria were peer-reviewed articles, only English language, and any publication year was accepted. Technical Standards were excluded. When possible, the searches were limited to the metadata of the articles to obtain only the articles that had the subject as a central focus point. On ACM Digital Library and SpringerLink, the searches were conducted on the full texts. All the hits were examined in a first round of sorting where abstracts were read to determine whether an article was relevant. During this process, the inclusion criteria were that the article included an explanation or definition of what Proactive HRI is, if they claimed to implement Proactive HRI, or if the term “proactive” was used in relation to, in the context of, or describing aspects of HRI. Figure 1 illustrates the date of the searches as well as the amount of hits and the number of relevant entries on the different databases. All searches were carried out in Autumn 2020. A free text search on Google Scholar was conducted as a last check to ensure that the most relevant papers were considered. The first 10 pages of hits were examined, and relevant articles that had not yet been found via the databases were included. The identified relevant articles were combined, and duplicates were resolved in Mendeley. The total amount of different articles found in the block search was 200 articles published in the year span 2001–2020. All papers of relevance that complied with the inclusion criteria available at the time of the search were included to have as much data and as exhaustive a search for this topic as possible. A procedure inspired by the Grounded Theory method (open coding and axial coding) [186] was used for the analysis of the found literature. The found articles have all been systematically reviewed according to a previously specified set of themes and elements: motivation/inspiration, application context, key methods, contributions, metrics, and so on. This was done through an iterative process where codes for some of the elements were formed and updated throughout the process.
Fig. 2.
Fig. 2. The distribution of the articles across publication years.
To gain a more elaborate overview of the development of the general field of Proactive HRI, an analysis of the identified articles has been conducted. Figure 2 illustrates the articles plotted according to their publication year in a bar graph. It can be concluded that the field of Proactive HRI indeed is a rather young, however, growing field with most of the articles published after 2010 and with an overall increase in the amount of published works per year.
4 Defining “Proactivity” in HRI
Through inspection of the corpus of literature, it becomes clear that there are large variations in how clearly different works define what they mean by “proactive.” Furthermore, it is evident that there are variations in the use of the term and what authors consider Proactive HRI. Through examination of the explanations provided in some of the works that attempt to clarify what they understand by “proactive” or “proactivity” in the context of the authors’ research, a valuable insight into the variations was attained.
An example of a work that includes a clear definition is the work by Luria et al. [109]. Although the work is on the boundaries of the scope of this review with respect to HRI, the article provides a clear definition of “Proactivity” in the context of their work as “initiation of interaction with a user.” Schmid et al. [153] is less direct in their wording but ascribes the term “proactive” to behavior where initiative is taken to carry out an action that triggers clarifying behavior by the opposite interaction partner. A similar yet less specific meaning of the term is described by Fink et al. [49], who describe a system that initiates interaction as proactive. Cámara et al. [26] describe proactiveness decoupled from interaction but as an agent’s ability to “exhibit goal-directed behavior by taking initiative,” thus, also ascribing to the connection between proactivity and initiative-taking. Jin et al. [79] provide a slightly different angle by stating that actions can be defined as proactive if they have the ability of creating or controlling a situation. Here, the focus is not only on the robot as the initiator, but also on the purposefulness of the robot actions. Pandey et al. [136], oppositely, do not focus on the ability to control a situation but rather the ability to contribute and support as proactivity when they define proactive behavior as “taking the initiative whenever necessary to support the ongoing interaction/task.” In the work by Sirithunge et al. [166], two specifications are provided. First, they set robot-initiated and proactive HRI alike and describe both as interaction where the robot is responsible for initiating the interaction. Second, proactive robots are defined as “Robots which identify the requirement of a certain situation and acts instantly without any instructions from outside.” The latter definition points towards an understanding that proactive robots have anticipatory capabilities. This notion is also expressed in the definitions and use of the term “proactive” in other works. An example hereof is a reference made by Peng et al. [140] to proactivity in occupational psychology from which they derive the following definition of proactivity for service robots: “The anticipatory action that robots initiate to impact themselves and/or others.” Chinchali et al. [33] likewise hint at the anticipatory understanding of “proactivity” when they describe how robots both need to proactively decode human intentions and, subsequently, utilize this knowledge for collaborative task satisfaction to do proactive decision-making. Likewise, Mok [121] clearly connects anticipation to proactivity when he states: “We defined a reactive robot to be one that responded and executed actions according to the user’s gestures, while a proactive robot anticipated a user’s impending needs and initiated an action to complete a task.” Another definition encountered states that “proactivity is acting in advance of a future situation, rather than just reacting” [90].
As can be seen from the presented selection of definitions and uses encountered in articles, there are some agreements on the meaning of “proactive” among the works. Several works express proactivity as taking initiative or controlling a situation rather than being reactive, while another group of works seems to lay out a clear relation between proactivity, anticipation, and acting based on estimations and assumptions. These differences in use of the term might partially stem from a discrepancy between the British and the American dictionaries regarding the word “proactive.” For instance, the American Merriam-Webster describes the meaning of proactive as “acting in anticipation of future problems, needs or changes” [118], whereas the British Cambridge Dictionary gives the definition “taking action by causing change and not only reacting to change when it happens” [27]. Though there are some similar elements to the two definitions, there are also differences that, again, also are identified in the use of the term in the corpus of literature. While some works can be clearly associated to one understanding or the other, there are also a wide range of works that are seemingly vague in their use of the term or portray recognition and acceptance of aspects of both initiative and anticipation. An example is the mentioned work by Peng et al. [140], who describe how their definition of robot proactivity indicates three key elements, namely, anticipation, initiation of action, and target of impact. They go on to note how initiation of action is about robot autonomy, thus, hinting at robot control. Tan et al. [173] discuss proactive behavior in relation to autonomy and specify that the difference of proactivity from mere autonomy is the anticipation of human intention. In these cases, one could argue that both anticipation and initiative play a role. Consequently, it can be noted that although some works can be ascribed to one understanding or the other, proactivity in HRI seems to be a continuum ranging from pure initiative to anticipatory skills. Based on this observation, it is difficult to give a single clear-cut definition of Proactive HRI. Instead, Proactive HRI can be described as HRI where the robot exhibits some level of either anticipatory or situation controlling capabilities. The two understandings contained in the provided definition of Proactive HRI can be used as a basis for classifying works. Nonetheless, the works can be further categorized based on other observed factors such as research or functional focus. These findings serve as the basis for the taxonomy proposed in Section 5.
A taxonomy is proposed to emphasize the nuances in Proactive HRI and provide a tool for specifying the focus of future works within this field. Our taxonomy is formed from observations made based on the identified corpus of literature and the use of the term “proactive” with respect to the work conducted within HRI. Please refer to Figure 3 for a selection of representative examples of the systems presented in the corpus. For readability, the taxonomy will be outlined herein, while relevant literature within each category will be provided in their respective subsections. Please see Figure 4 for an overview of the proposed taxonomy.
Fig. 4.
Fig. 4. The proposed taxonomy based on the literature review. The number in each box indicates the amount of articles that grounds the category. Note that not all of the 200 articles found contribute to a category, and that a single paper can contain several different components and, hence, be included in multiple subcategories.
A basic observed distinction present in the corpus is the question of which agent is to exhibit proactivity. While most of the works focus on endowing a robot with proactive abilities, a minority of works focus on human proactivity and facilitating such behavior through HRI or other means during HRI. For either agent, human or robot, the proactivity can be distinguished based on the two main understandings of proactivity mentioned in Section 4. In relation to robots, this results in the two main branches Anticipatory Robot Behavior and Robot Initiative. Furthermore, a small selection of works that do not conform well with this classification introduce a Miscellaneous category, which will be further elaborated in Section 5.3.
The two branches are subcategorized based on functional focus of the technologies or investigations presented in the analyzed literature. With respect to Anticipatory Robot Behavior, this gives rise to several subcategories of commonly researched topics that overall relate to a pipeline of identifying intentions of the human or other inferred factors and, subsequently, initiate the adequate robotic actions. As indicated by the two large color blocks within this branch in Figure 4, these topics in general aim to endow robots with anticipatory abilities to determine and plan either necessary robot motion or robot actions or tasks on a more abstract level. Some of the subcategories presented within this branch have relevance for the fulfillment of both objectives while others are specific for one of them. Additional details are given in Section 5.1.
The branch of Robot Initiative contains the three subcategories Robot-Initiated Interaction, Robot-Initiated Actions, and Robot-Controlled Influence. While these three categories might not always be mutually exclusive in presented systems, there are some clear differences in the functional purposes presented in these categories. These are further explained in Section 5.2.
The branches and subcategories introduced in the taxonomy are substantiated through examples in the literature. The main themes of each branch and subcategory identified in the literature are elaborated and exemplified through a select few relevant articles related in the dedicated sections below. In this fashion, the state-of-the-art within each research topic is outlined.
5.1 Anticipatory Robot Behavior
As the most ramified of the branches in the taxonomy, the works included in Anticipatory Robot Behavior focus on a range of research topics related to endowing robots with anticipatory capabilities. Though the foci vary, the close correlation of the topics often cause several of them to be addressed in individual publications. Figure 5 provides an overview of all the articles that relate to each of the research topics represented in the subcategories.
5.1.1 Anticipatory Inference.
As anticipation is about acting in advance of a future state or situation, and, in the context of HRI, without explicit input from a human partner; estimating or predicting relevant, possibly “hidden,” factors or information is of central importance. This category covers a range of methods all centered around inferring or estimating such unknown data, which serves as the basis for anticipatory proactive robot behavior. The desired information varies in nature. For anticipation in relation to motions, the information estimated might be humans’ intended trajectories, space occupancy, risk, perturbations, and so on. In the context of actions and tasks, examples of the desired information include human intentions, emotions, wants, needs, goals, human skill levels, human actions, and subsequent robot actions directly. Similarly, many and variate terms are used to describe the disciplines covered in this category dependent on the aim: motion prediction, intention inference, human intention recognition, risk estimation, and so on. The earliest work in the corpus addressing this challenge is by Breazeal et al. [20]. Their approach builds on Simulation Theory and the idea that we as humans use our own responses to simulate others’ minds to infer their mental states. Utilizing this idea, Reference [20] presents a robot that infers human intentions based on a mapping between observed human movements and a library of known robotic movements and their corresponding end goals, assuming the found goal mirrors the human intent. A limitation of this method is the fixed finite number of interpretable goals by the robot.
A different approach is to represent each human intention by their corresponding action sequences in the form of Finite State Machines (FSMs). By comparing an observed action sequence to the actions sequence of the FSMs of a known intention, a probabilistic weight is obtained that indicates how closely the intention represented by the FSM relates to the actual current intention [9].
The works by Breazeal et al. [20] and Awais and Henrich [9] are concerned with deducing some higher-level abstract premises, in this case intentions deduced from measurements or cues, that are of importance to the subsequent robot action selection. However, other works take on a more direct strategy where robot actions are inferred by, e.g., a known behavioral pattern or a sequence of steps that must be carried out to complete a collaborative task [90].
Fig. 5.
Fig. 5. The list of articles that form the basis for the subcategories in Anticipatory Robot Behavior. The illustrations depict examples of the principles in the subcategories. “Other Factors” under Anticipatory Inference covers skills, emotions, and so on. In Contextual Awareness, “Other Context” includes research where the types of contextual information considered are unclear.
5.1.2 Contextual Awareness.
Contextual Awareness is typically addressed in relation to implementation of anticipatory inference. It can be including context to predict human motion behavior or to support the deduction of human intentions, and so on. While works focused on prediction or inference of “hidden” or future factors do so based on different kinds of data, some works explicitly include additional information that is not considered central to the factor of interest. These are the cases that portray a level of contextual awareness. Liu et al. [106] proposed a method for intent recognition by domestic assistive robots that combines object affordances and context features related to intentions for the different objects. The objects cup and window are considered where the related context features can be hot weather, low drinking frequency, medicine taking time, and nice weather outside, poor weather outside, quiet outside, noisy outside, respectively. The intentions are recognized using a Bayesian network-based intention recognition model that considers the combination of this information and observations of contextual features. In a later work, a Fuzzy-Naïve Bayesian Network algorithm has been adopted for the same purpose [105].
5.1.3 Ambiguity Resolution.
Some works that focus on anticipatory inference also explore how a robot handles, e.g., ambiguous intentions. This includes how the robot is to act under such circumstances, but also policies for how the ambiguity can be resolved, and a single intention can be confirmed as the dominant one. Though this is a valuable functionality for Proactive HRI, only a few works have attended to this challenge. The select group of authors behind the majority of the works propose different approaches dependent on the amount of likely human intentions identified. The methods proposed utilize what the authors call “proactive execution,” where an action is executed under uncertainty in case only a few human intentions are likely. Besides being the “best guess,” the purpose of the proactive execution is to gain more information and determine the “true” human intention [153, 154, 156, 189].
5.1.4 Task Selection and Planning.
Task Selection and Planning is closely related to the general Anticipatory Inference and is in some cases inseparable from that category when it comes to task selection. Nonetheless, there are examples of works that present a clear two-stage approach where, e.g., estimates of intentions are obtained in the first stage and then the appropriate robot actions are selected in a second stage based on the estimated intentions. Furthermore, while the category of Anticipatory Inference mainly focuses on finding the basis for determining which action should be carried out, a range of works consider methods for determining when the adequate action should be carried out. This is a question of action or task planning, rather than mere task selection. The question of task planning or timing of the execution of robotic tasks or actions has been mentioned in 8 of the 33 articles in this category. Of the 8 articles, 5 of them are consecutive articles by the same authors, who first propose the use of a probabilistic temporal prediction method based on causal probability and eventually advance a hybrid temporal Bayesian network to a composite node temporal Bayesian network for the prediction of the best robotic actions as well as the best timing of those actions [90–94].
Some of the above-mentioned works include the ability to determine which robot action is adequate. This is also explored in the work by Schmid et al. [153], where actions are selected based on the most probable intentions in an intention estimate and are subsequently filtered using a proposed method inspired by Lorentz’s “psycho-hydraulic model” [107]. Another example is the work by Pinheiro and Bicho [143], where a control architecture consisting of coupled Dynamic Neural Fields is proposed for determining goals and executing the appropriate action in an assistive robot. Here, specifically one of the layers in the architecture, the Action Execution Layer, is responsible for selecting the most appropriate behavior. In a more recent work, Wojtak et al. [190], likewise, propose a method based on DNFs that offers temporal integration of inputs to learn sequences and, thus, enable proactive action selection.
5.1.5 Motion Planning.
The Motion Planning category builds on a selection of papers that use proactive anticipatory skills to improve the planning of robotic motion in relation to HRI. This can be motion planning in 3D space, e.g., for a robot manipulator or the 2D motion planning for a mobile robot navigating an environment.
An example of a 2D application is the use of anticipated human trajectories to determine the appropriate approach for social robots [81, 177, 178].
One of the themes handled in anticipatory motion planning in 3D are handovers during HRC. Nemlekar et al. [129] propose a method for anticipating the Object Transfer Point (OTP) for handovers from humans to robots. The method calculates an offline OTP based on studied human preferences, and a dynamic OTP and combines them to achieve better accuracy. Another interesting theme is collaborative object transportation, for which presented methods anticipate human intended motions and plan robotic motions accordingly [25, 161]. Similarly, anticipatory motion planning is investigated as a solution in pHRI, where the anticipation of human intentions is used to predict trajectories to improve load sharing. This is relevant in general HRC [85, 112, 131] but also in relation to exoskeletons [75, 76], where it can facilitate compliance of the wearable robotics with a reduced interaction force.
5.1.6 Collision Avoidance.
This category is closely related to the research topic of motion planning and might be considered a special case of that category. However, since the topic of collision avoidance is not the only aspect considered in motion planning, it is presented separately. This category also contains examples of both 2D and 3D applications.
The examples of collision avoidance in 2D are found in relation to mobile robots and autonomous vehicles. Leung et al. [100] propose a traffic-weaving method for autonomous vehicles, which includes a prediction model of human behavior to find collision-free trajectories. The works related to mobile robots are generally concerned with avoiding collision while navigating human-dominated environments [84, 177, 178]. The works aim at proactively avoiding collisions in these environments by considering human motion as well as social constraints.
The works presenting methods related to collision avoidance in 3D either utilize anticipated motion [68, 134] or some representation of risk [151, 152, 172]. The methods are anticipatory of a general nature but can be considered Proactive HRI due to the application in relation to interaction with humans. The aim of the methods is to prevent collisions during close proximity HRI such as during HRC.
5.2 Robot Initiative
The branch of Robot Initiative is based on the selection of works that are focused on the ability of the robot to take control and initiative to establish interaction, initiate actions during interaction, or create or alter situations proactively according to monitored factors or internal goals. Figure 6 provides an overview of the articles related to each of the subcategories within this branch.
5.2.1 Robot-Initiated Interaction.
A group of works are focused on how a robot can initiate interaction. More specifically, the focus is on the robot actively carrying out interventions to seek out or request interaction with a human. This topic is typically researched in relation to social or service robots, as these can be situated in dynamic environments where the humans might not initiate interaction themselves.
Fig. 6.
Fig. 6. The list of of articles that form the basis for the subcategories in Robot Initiative. The illustrations depict examples of the principles in the subcategories. “Other Interaction Initiation” includes papers focused on, e.g., dialog and interaction initiation in general. “Other initiated actions” includes general research of robots initiating actions.
Applications presented in the works within this category vary in respect to who is the beneficiary. In some works, the purpose of initiating interaction is to provide the human with a service, while other works investigate how to establish interaction with the purpose of the human assisting the robot. In the cases where a human is the beneficiary, the robot might initiate interaction to provide additional information about exhibitions in a museum [77, 144–146] or offer decision-making support, e.g., in relation to a potential purchase [104, 140]. Rosenthal et al. [149] present a work where a robot navigating an office space is the beneficiary. The work presents a model for seeking out occupants in the environment to request assistance for the completion of the robot’s task.
The works in this category cover themes such as how a mobile robot should approach potential users, how different levels of proactivity are perceived, how a robot should initiate interaction, and so on.
5.2.2 Robot-Initiated Actions.
Works in this category focus on the initiation of robotic actions. Whereas Robot-Initiated Interaction is concerned with robots establishing interaction where there previously was none, this category of works is more focused on how and when robots should initiate actions in the context of ongoing HRI. Naturally, this category is closely linked to the research topics in the branch of Anticipatory Robot Behavior, as robot-initiated actions are realized by means of the methods presented in those topics. However, this group of works is focused on the robot initiative and the effects of this behavior rather than how this behavior is achieved. Examples of works in this category include the work by Baraglia et al. [13], who proposed and compared three initiative models for a robot collaborating with a human partner on a tabletop manipulation task. Tan et al. [173] present their work with a slightly different aim, namely, to investigate the relationship between proactively initiated actions by a social robot and the perceived anthropomorphism of the robot. In their work, they carried out Wizard-of-Oz experiments where subjects interacted with the robot with five different levels of robot proactivity to determine the effects of the proactivity level.
5.2.3 Robot-Controlled Influence.
This category covers examples of robots enforcing their own “intentions” during HRI with the purpose of actively influencing the situation or circumstances of the interaction in a certain way. This can be to improve certain aspects of the interaction or to expedite the robot’s “internal goals” or “interests.” The work by Gray and Breazeal [63] is an example of what could be perceived as a robot taking initiative to facilitate achievement of its autonomous internal goals during HRI. In this work, the correlation between action perceived visually by humans and the change in mental states is learned. This information is then utilized by the robot to manipulate the human mental state for its own advantage. A different example is presented by Lazar et al. [95] in their work related to autonomous vehicles, where they investigate how robot-controlled influence through planned interactions with human-driven cars can be used, purposefully, to improve road configurations. This could be done with the goal of facilitating positive influence on social objectives, e.g., fuel consumption.
5.3 Miscellaneous
Some works in the corpus simply have a too vague usage of the term proactive in relation to their work to give merit to a placement in one of the suggested categories [8, 32, 62, 126, 184], the notion they convey seems to be not proactive at all [47], or the relation between the proactive behavior and HRI is questionable [142, 168]. An example of the latter is the work by Cámara et al. [26] that addresses robot proactivity as a characteristic of self-aware computing systems but not specifically in the context of HRI. Such works have been grouped in Miscellaneous. Some of these works could, arguably, have been excluded from the corpus altogether but have been included due to the criteria set for the structured sorting process.
5.4 Human Proactivity
This category consists of a group of works concerned with handling, inducing, or facilitating proactive behavior by humans through HRI. With a different focus from most works (endow robots with human-like proactivity; see Section 6.2), this category can be viewed as an outlier to the general tendencies in research within Proactive HRI.
Examples of works that fall into this category are works concerned with people struggling with social shortcomings, such as children with Autism Spectrum Disorders (ASD). The work by Feil-Seifer and Matarić [46] focuses on development of systems for interventions for children with ASD to improve on social deficiencies. More specifically, addressing reduced self-initiation using robots is suggested, and the effect of a robot on the proactive social interaction is observed. François et al. [52] conducted similar work investigating the effects of robot-assisted play for children with ASD with the purpose of encouraging proactivity and initiative-taking. Robins et al. [148] hypothesized that repetitive exposure of children with ASD to a small interactive humanoid robot will increase basic social interaction skills. They observed a more social response to a human dressed and acting as a robot compared to a human and, subsequently, discovered that repetitive exposure to a small humanoid robot facilitated the emergence of spontaneous and proactive interactions.
Exemplifying differently are works that strive to encourage humans to be proactive against potential dangers during HRI or HRC. In Reference [115], two classes of techniques for ensuring safe HRC are described, one of which is focused on cultivating proactive human behavior by raising awareness through cognitive aids. In the article, these techniques are investigated in a Virtual Reality (VR) environment where the cognitive aids during HRC are implemented as audio and visual cues and continuous adaption of these based on environmental conditions.
Yet other works deal with how some proactive human behaviors can be a beneficial element in HRI [158, 159] or how unfavorable human proactivity can be handled [71]. Figure 7 provides a condensed overview of the articles in this category.
Fig. 7.
Fig. 7. The list of articles that forms the basis for the subcategory Human Proactivity. The illustrations depict examples of the principles in the subcategories.
6 Discussion
Reflections related to the proposed taxonomy, as well as a selection of noteworthy observations made based on general tendencies in the analyzed literature, are next discussed.
6.1 Discussion of Literature Search and Taxonomy
During the literature search process, the focus in the block search on open terms directly describing interaction scenarios has consequences for the outcome. A test search conducted to examine these consequences including more field-specific terms has indicated that only the inclusion of “autonomous vehicles” gave a higher yield. In a future taxonomy, this term could be included. A natural consequence of the search process is also that only works that relate proactivity to their research topics are obtained, and more literature exists on, e.g., intent recognition, which otherwise is closely related to proactive behavior, than found in this study. Similarly, sorting the hits based on metadata only means that works are excluded if they only address proactivity in the full text.
The field of Proactive HRI is a field interwoven with elements from several disciplines and where topics interconnect. The process of defining a taxonomy in relation to Proactive HRI is not straightforward and has a series of related challenges. As is indicated in Section 4, giving a single unanimous definition of proactivity in relation to HRI is even a challenge because of the continuum nature of the meaning and use of the term. It should be noted that this also affects the proposed taxonomy. While it would be preferable to be able to assign a presented method resolutely to a single category, the reality is that works often engage with research topics from several of the categories of the taxonomy simultaneously. Even more so, some works clearly present topics that contain elements from both understandings of proactivity, pointing at the continuous and interconnected nature of the spectrum, which can hardly be portrayed to its full extent in a categorical taxonomy.
The work by Rashed et al. [145, 146] is an example hereof. While the main focus is to endow a museum guide robot with the ability to proactively approach visitors and initiate interaction by providing extra information about an exhibition piece, the work also depends on anticipatory aspects of estimating the directed attention of the visitor as well as the interest in additional information about the piece.
Although works might not fit solely into a single subcategory in the taxonomy, it is the strong belief of the authors that the taxonomy can serve as a valuable tool for clarifying the intent and the focus points of a specific work.
Another, challenge when forming a taxonomy based on the corpus of literature is the differences in the aim of the works. While some works focus on the technical methods and challenges in relation to endowing a robot with the desired proactive abilities, other works are more concerned with evaluating the effects of the proactive robot behavior. This means that a category such as Anticipatory Robot Behavior contains works concerned with enabling the prediction of, e.g., needs and corresponding supportive robot actions as well as works investigating whether anticipatory behavior is desired by human partners. These investigations are, for example, carried out using Wizard-of-Oz experiments simulating the proactive robot behavior.
In relation to the definition provided in Section 4, it is important to note that even with an agreed characterization of the term, determining whether a specific work can be considered Proactive HRI can be a matter of interpretation. One of the commonly presented notions of proactive behavior is that it stands in contrast to reactive behavior. Robots taking initiative might clearly comply with this notion; however, when it comes to anticipatory robot behavior, the lines become more blurred. In practice, anticipatory behavior might be achieved by acting according to predictions, or it might be achieved by action upon early cues or precursors of certain events. With respect to predictions, it may be clear that this practice is in fact not simply reactive; but for action upon cues, one might argue that this is simply “early reaction” or an adjustment of what is the triggering observation.
Another discussion point is whether robotic actions carried out on own initiative with no anticipation or adjustment to the situation in fact can be considered proactive or simply a case of autonomous behavior. The work by Huang et al. [73] serves as an example of such a discussable case. In their work, they propose a “proactive coordination” method for human–robot handovers. The description provided for this coordination method explicitly states that the robot does not consider the user’s task demand and simply proactively fetches a new item as soon as it has an empty gripper. In this context, the word “proactively” seemingly could be a synonym for either “autonomously” or “in advance” and seems to describe a behavior that is automatic rather than “reflective” or intelligent in the way most proactive behaviors appear to be.
6.2 Motivation behind Proactive HRI
As is custom in general, many of the included articles start off by motivating their work. While some are focused on the background for their novel technical ideas, a selection of articles argues for the general need for robot proactivity. By examining these arguments, an insight into the line of thought and assumed benefits motivating the introduction of proactive abilities to the robot in HRI can be gained. These motivations will be addressed in relation to the two main understandings/branches identified in the taxonomy.
6.2.1 Source of Inspiration.
Works addressing topics related to anticipatory robotic behavior often argue for this behavior based on observations from human-human interaction. Awais and Henrich [9] point out the importance of proactivity for effective cooperation, stating how humans collaborating on a task are required to be proactive toward each other to secure intuitive collaboration. They continue, “It is equally important in the human-robot interaction that the robot should be proactive according to the collaborating human depending on the current situation.” Similarly, Kwon and Suh [94] mention how interacting humans predict each other’s intentions and synchronize their reactions accordingly with the benefit of ensuring seamless interactions through proactive responses. It is further argued that delays in human–robot interactions caused by frequently occurring rigid request-and-react patterns can be reduced and more seamless interaction be facilitated by robots’ abilities to predict future events based on observed human activities.
Similarly, articles engaged with robot initiative often point to human behavior for the inspirational source of the behavior they strive to establish in robots. In Reference [2] the desired proactive behavior of initiating actions for humans without prior interaction is compared to the spontaneous behavior people exhibit.
Some works design proactive robotic behavior in HRI based on inspiration gathered by studying human-human interaction directly [73, 199].
6.2.2 Argued Advantages.
The main advantages argued for in relation to Proactive HRI are intuitiveness and understandable or natural interaction seen from the human partners’ point of view [136, 153]. Another important point is the argumentation related to efficiency. This is expressed in different ways in the literature; in some cases mentioned as fluency and in other cases, more concretely, as the minimization of delays or idle time for either partner [110, 199]. Finally, some works argue for more specific advantages such as, e.g., the ability to efficiently execute tasks while avoiding collision [68] or the benefit for the human in being freed from supervising and commanding a proactive robot [3].
6.3 Common Applications in Proactive HRI
As a large part of the motivational arguments provided in Section 6.2 suggest that a desirable benefit of robot proactivity in HRI is the more human-like interaction, it is expectable that robot proactivity is applied in contexts with many human–robot interactions and in contexts where the robot operates in human-centered environments. This is confirmed by the contexts and applications observed within the Proactive HRI corpus. Figure 8 reports findings related to applications and interaction characteristics. The classification draws inspiration from Reference [41]. A large amount of the applications are social robots, and more specifically service, assistive and domestic robots navigating in human centered environments. An example hereof is the assistive robot developed in the RAMCIP project, which is focused on assisting elderly in daily living activities with increasing levels according to the elderly’s level of impairment [48]. The idea is that the robot proactively adjusts the level of assistance offered in accordance with the need based on the estimated skill level of the elderly. Another example is the museum guide robot proposed by Iio et al. [77], which proactively offers information about the displayed exhibitions based on the human’s estimated focus of interest. Truong and Ngo [177] present a work focused on a service robot situated in a crowded human-ruled environment, where the challenge is to navigate safely and sociably. For this purpose, they propose a proactive social motion model used for proactively adjusting the robot’s navigation according to social cues and anticipated trajectories.
Fig. 8.
Fig. 8. Overview of applications and interaction characteristics, i.e., devices, relationships, and beneficiaries. The numbers in the parentheses indicate the number of works related to the application. The square brackets indicate representative example references for the different groupings.
Another common field in which Proactive HRI is applied and investigated is HRC, where a large number of interactions and a still rising need for flexibility of the robot calls for more analytical, adaptive capabilities in the collaborative robots. Robots can collaborate with humans in different ways under different conditions. Tabletop HRC is a condition often explored in articles. Schulz et al. [157], e.g., investigates the human experience of different robot behaviors, proactive among others, during a collaborative tabletop building task.
These two application areas are the dominating ones within the corpus. Nonetheless, other areas such as the aforementioned “autonomous vehicles” [71, 95], “wearable robotics” [19, 75, 76], “robotic objects” (“RObjects”) [49, 121, 123], etc. are represented by a few examples.
6.4 Measurements and Evaluation in Proactive HRI
As mentioned earlier, Proactive HRI is a field that is characterized by the intersection of multiple disciplines. In the corpus of literature, this is reflected by the wide variety of focuses and aims of the works. Some works propose general concepts or frameworks, other works are focused on technical development and in some cases evaluation of the implemented systems, and yet again another group of works investigates the psychological effects of different robotic behaviors. One difficulty that arises from this variety is that it is hard to compare works and results across the field. Dependent on the aim of the work and which scientific discipline the work takes its onset in, different methods with each their own measures are used to obtain and evaluate the findings.
Works addressing prediction or recognition of factors usually evaluate their performance with traditional machine learning and deep learning measures such as confusion matrices [45, 128], recall and precision, or measures based hereon [30, 119, 199].
For evaluating the effect and efficiency or performance of proposed systems yet another set of metrics are used. Here, measures such as completion time, idle times and number of actions carried out by the robot and human, respectively, during a collaborative task are considered [13, 44, 196].
Other works use metrics or elements for their evaluation that are very specific to their works. An example is the interaction diagrams used by Moulin-Frier et al. [125] as a visualization of the interactions between test subjects and their proactive robot to portray the processes, their time of occurrence, and their duration. The road capacity measure mentioned in Reference [95] is, similarly, a specific measure used to evaluate how successful the proposed strategy for proactively influencing humans to optimize conditions.
One could argue that this is a necessary evil of the field of Proactive HRI due to its wide spectrum and that different focal points naturally call for different evaluation methods and measures. However, even works with the same disciplinary focus sometimes prove hard to compare due to individualized and handcrafted measures. An example of this issue is the different studies investigating attitudes towards robot proactivity that use Likert scales to evaluate different measures. Some studies utilize existing questionnaires or scales directly or in modified versions, while others entirely compose their own set of scales. Mok et al. [123] evaluated subjects’ perceptions of interaction with a system, respectively, with or without proactive and expressiveness by using an adapted version of Dillard’s Relational Message Scale that consists of a set of 7-point Liket scales addressing different relational factors: immediacy, affect, similarity, dominance, among others. In contrast, Zhang et al. [196] and Schulz et al. [157] both evaluated the human experience of interaction using self-compiled, non-standard 7-point Likert-scale questionnaires.
6.5 Human Perception of Robot Proactivity
With a widespread advocation for the desirability of robot proactivity with the purpose of improving HRI, it is relevant to investigate how robot proactivity influences the interaction and the users and examine how robot proactivity is perceived by humans. In the literature, investigative works are found that address these topics. As expected, observations with respect to these topics may be closely connected to the kind of robot proactivity and the context of application of the robot and vary accordingly.
The improved efficiency that is used as an argument for implementing proactive behaviors has in fact been confirmed by several works. Anticipatory abilities have been proven to reduce waiting time in a collaborative assembly task [110], reduce reaction time in human–robot handovers [129], and reduce completion time [196].
Baraglia et al. [13] found in their study that human collaboration with a proactive robot resulted in high ratings of the system and better team fluency compared to collaboration with a reactive robot. Zhang et al. [196] have investigated whether proactive supportive actions based on inferred human goals and intentions carried out by a robot during a collaborative urban search and rescue task are desirable. They found that subjects preferred the robot with the ability to provide proactive support compared to one without this ability. However, their results also suggested that the proactive support led to a higher cognitive load.
Although many works present results that argue that robot proactivity is desirable or preferable compared to other reference behaviors, some studies also highlight that proactivity can have negative effects. Humans may have reservations toward proactive robotic behavior. The results obtained by Deutsch et al. [39] in their study of older adults’ attitudes toward proactivity as a design aspect for home robotic devices revealed that the majority of the utterances collected related to proactivity were negative. Test subjects expressed concerns about the influence on their own activity level, the degeneration of their abilities, and the loss of independence and control; and some stated that they preferred explicitly telling the robot what to do rather than it taking initiative. Cesta et al. [31] studied elderly peoples’ attitudes toward a robot mediator offering cognitive support. The robot, which was evaluated offering both on-demand and proactive support in relation to a specific set of daily living scenarios, was overall reported useful and indicated as an acceptable solution by the test subjects. Furthermore, the preference for the robot’s support and its usefulness was significantly higher when it offered proactive support related to emergency and healthcare situations compared to on-demand and proactive suggestions.
The study by Peng et al. [140], which investigated the effects of different levels of robot proactivity on user perception during a decision-making support process, found that the medium level of proactivity was more helpful and preferable compared to the highest level of proactivity. Kraus et al. [87] investigated the effects of proactive dialog strategies on the human trust toward the system. The level of proactivity was varied through the degree of initiative or freedom to actively influence a given situation. They introduce three levels of proactivity in the robot’s behavior in addition to a reactive baseline behavior. The three levels of proactive behavior are “Notification,” simply notifying the user that the robot has a solution to offer to the problem at hand, “Suggestion,” actively providing information on a solution, and “Intervention,” take control and solve the problem using the found solution. As in Reference [140], they found that the medium level proactive strategies were preferred. In both Reference [140] and Kraus et al. [87], the medium levels of proactivity are characterized by the robot proactively offering intelligent services to the user but requiring the user to take action to utilize the robot’s service, thus maintaining some level of user control. In contrast to [87, 196] found that proactive system strategies, in their experiments, did not result in a significant increase in cognitive load.
To summarize, both the effects and the perception or desirability of proactivity in HRI vary according to the application, the way the proactive behavior is realized, and who the user of the system is. As such, it can be concluded that the benefit of proactive robotic behavior is not absolute, and that more research is needed to fully understand the factors influencing the success of proactive robot behavior.
7 Open Research Questions
With Proactive HRI being connected to many disciplines, there is rich potential for further research in several directions. The studies related to proactivity in children with ASD reveal that robots have potential to facilitate proactive behaviors. The study by Robins et al. [148] further suggests that the appearance of the robot has influence, and the studies of this connection between factors such as appearance and the evoked proactivity could be further investigated.
Similarly, numerous works have studied robot proactivity and investigated how a robot can anticipate or recognize a need and act accordingly. However, shortcomings are the rigid context-specific programming of the robot behaviors, which also requires further investigation. True potential lies within flexible solutions that are able to learn connections between observations and feasible robot contributions. This also emphasizes the need for contextual inclusion and the ability to autonomously extract and weigh relevant contextual information from, possibly, several modalities. This could support further explorations of how ambiguity can be resolved in proactive robot behavior.
The effects of interaction between humans and proactive robots also hold potential for further discovery. Though researchers have investigated human preference of robot proactivity, there are still insights to be gained from evaluating real experienced interactions between a subject and a robot. The results in both References [31, 39] are based on elderly people observing video materials and providing answers in interviews and questionnaires. Theoretical and video sample-based studies play important roles; however, measuring subjects’ experiences, opinions, and preferences through evaluations obtained in the context of as realistic scenarios as possible can improve ecological validity. As such, for a greater understanding of in situ experiences, HRI researchers are encouraged to conduct more studies with real robots like in Reference [44], measuring, e.g., elderly peoples’ opinions towards assistive robots. Furthermore, it should be noted that preference of robot proactivity might be one thing when discussed in theoretical scenarios, something else when experienced in a real encounter with a novel and unknown system, and something entirely different when experienced throughout long-term practical use. Thus, for gaining beneficial knowledge of these relationships and the preferences in long-term interactions, more longitudinal studies would be valuable.
The results presented in both References [39] and [31], furthermore, highlight that not all robot proactivity is equally desired. Consequently, the need for more research into when the robot should or should not engage, e.g., to support human abilities and autonomy [86], and methods that enable robots to do such reasoning correctly, is apparent. In general, extensive studies of the effects and desirability of proactive robot behavior is lacking. Although the current research indicates that there are differences between applications, levels of proactivity, the types of proactivity, user profile, and so on, too little research has been conducted that is fine-grained enough to map out the correlation between these variables and the resulting effects. Investigating different proactive behaviors, rather than a single proposed proactive behavior, and comparing to a baseline could also prove valuable to offer an improved, more fine-grained insight into what is positive and what has negative consequences. This will decrease the likelihood of robot proactivity being falsely rejected in its entirety due to an inappropriate design.
While it is argued that robots working more independently and proactively will relieve humans and reduce their cognitive load [125], investigation suggests that this behavior in human–robot teaming increases the cognitive load [196]. This indicates that there could be a discrepancy between the expected and actual effect on cognitive load or that it differs between setups. With new robot abilities and behavior come new unknown effects on the users. One important research question is how does proactivity in, e.g., HRC, affect the assumptions made by the human partner about the robot’s intelligence. Can proactivity, for example, cause misplaced trust?
The vast majority of the works in the collected corpus focus on proactivity in a one-to-one human–robot interaction, which is expectable, as interaction in environments with multiple agents often end up as several one-to-one interactions. However, research into proactive behavior in multi-agent interaction with more than one human and/or robot is scarce. For this perspective, it could be interesting to expand the knowledge on the effect of robot proactivity on team dynamics or investigate the potential and effects of a collective proactivity by a robot team, e.g., in a human-dominated environment.
Finally, a great challenge in this diverse field is the large variety of measures used. As seen in other fields, commonly accepted metrics and benchmark datasets can ensure a rapid development. Also, objective measures for evaluating HRI can facilitate clear and precise results. A push in this direction is highly recommended.
8 Conclusion
Robots are to a still greater extent applied in fields where they interact with humans that require flexibility, adaptiveness, and autonomy. Consequently, robot proactivity in HRI has gained interest. In this article, the works related to Proactive Human–Robot Interaction (Proactive HRI) have been mapped out through a structured literature study. This work has provided an insight into Proactive HRI and its nuances and has given a definition of proactivity in the context of HRI. A taxonomy of research topics related to the field has been proposed that will serve as a tool for specifying the aim and the nature of future works as well as a map for researchers who are new to the field. Furthermore, the taxonomy is discussed as well as some general tendencies observed within the corpus of literature. Finally, an overview of challenges and open research questions related to proactive HRI has been provided.
References
[1]
Hojjat Abdollahi, Ali Mollahosseini, Josh T. Lane, and Mohammad H. Mahoor. 2017. A pilot study on using an intelligent life-like robot as a companion for elderly individuals with dementia and depression. In 2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids). IEEE, 541–546.
Okada Akiho and Midori Sugaya. 2016. Impression evaluation for active behavior of robot in human robot interaction. In International Conference on Human-Computer Interaction. Springer, 83–95.
Ignasi Andrés, Leliane Nunes de Barros, Denis D. Mauá, and Thiago D. Simão. 2018. When a Robot Reaches Out for Human Help. In Ibero-American Conference on Artificial Intelligence. Springer, 277–289.
Muhammad Usman Ashraf, Muhammad Awais, Muhammad Sheraz Arshad Malik, Ijaz Shoukat, and Muhammad Sher. 2017. Proactive intention-based safety through human location anticipation in HRI workspace. International Journal of Advanced Computer Science and Applications 8, 4 (2017), 378–384.
Muhammad Usman Ashraf, Muhammad Awais, Syed Hasnain Abbas Naqvi, Syed Muhammad Saqlain, Imran Khan, and Anwar Ghani. 2019. MDP-led Proactive Intention Recognition for Improved HRI Settings. International Journal of Computer Science and Network Security 19, 6 (JUN 302019), 91–97.
Ella M Atkins. 2007. Physically-proximal human-robot collaboration for air and space applications. In Proceedings of the 2007 Workshop on Performance Metrics for Intelligent Systems. 216–223.
Muhammad Awais and Dominik Henrich. 2012. Proactive premature intention estimation for intuitive human-robot collaboration. In 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems. IEEE, 4098–4103.
Ritta Baddoura and Gentiane Venture. 2013. Social vs. useful HRI: experiencing the familiar, perceiving the robot as a sociable partner and responding to its actions. International Journal of Social Robotics 5, 4 (2013), 529–547.
Ritta Baddoura and Gentiane Venture. 2014. Human motion characteristics in relation to feeling familiar or frightened during an announced short interaction with a proactive humanoid. Frontiers in Neurorobotics 8 (2014), 12.
Ritta Baddoura and Gentiane Venture. 2015. This robot is sociable: close-up on the gestures and measured motion of a human responding to a proactive robot. International Journal of Social Robotics 7, 4 (2015), 489–496.
Jimmy Baraglia, Maya Cakmak, Yukie Nagai, Rajesh Rao, and Minoru Asada. 2016. Initiative in robot assistance during collaborative task execution. In 2016 11th ACM/IEEE International Conference on Human-Robot Interaction (HRI). IEEE, 67–74.
Jimmy Baraglia, Maya Cakmak, Yukie Nagai, Rajesh P. N. Rao, and Minoru Asada. 2017. Efficient human-robot collaboration: when should a robot take initiative? The International Journal of Robotics Research 36, 5--7 (2017), 563–579.
Suna Bensch and Thomas Hellström. 2014. Towards proactive robot behavior based on incremental language analysis. In Proceedings of the 2014 Workshop on Multimodal, Multi-Party, Real-World Human-Robot Interaction. Association for Computing Machinery, 21–22.
Francesca Bianco and Dimitri Ognibene. 2019. Functional advantages of an adaptive Theory of Mind for robotics: a review of current architectures. In 2019 11th Computer Science and Electronic Engineering (CEEC). IEEE, 139–143.
Francesca Bianco and Dimitri Ognibene. 2019. Transferring Adaptive Theory of Mind to social robots: insights from developmental psychology to robotics. In International Conference on Social Robotics. Springer, 77–87.
Estela Bicho, Wolfram Erlhagen, Emanuel Sousa, Luis Louro, Nzoji Hipólito, E. C. Silva, Rui Silva, Flora Ferreira, Toni Machado, M. Hulstijn, Y. Maas, E. de Bruijn, R. H. Cuijpers, R. Newman-Norlund, H. van Schie, R. G. J. Meulenbroek, and H. Bekkering. 2012. The power of prediction: robots that read intentions. In 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems. IEEE, 5458–5459.
Baldin Llorens Bonilla and H. Harry Asada. 2014. A robot on the shoulder: Coordinated human-wearable robot control using coloured petri nets and partial least squares predictions. In 2014 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 119–125.
Cynthia Breazeal, J. Gray, G. Hoffman, and M. Berlin. 2004. Social robots: Beyond tools to partners. In RO-MAN 2004. 13th IEEE International Workshop on Robot and Human Interactive Communication (IEEE Catalog No. 04TH8759). IEEE, 551–556.
Connor Brooks, Madhur Atreya, and Daniel Szafir. 2018. Proactive Robot Assistants for Freeform Collaborative Tasks through Multimodal Recognition of Generic Subtasks. In 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 8567–8573.
Martin Buss, Daniel Carton, Barbara Gonsior, Kolja Kuehnlenz, Christian Landsiedel, Nikos Mitsou, Roderick de Nijs, Jakub Zlotowski, Stefan Sosnowski, Ewald Strasser, Manfred Tscheligi, Astrid Weiss, and Dirk Wollherr. 2011. Towards proactive human-robot interaction in human environments. In 2011 2nd International Conference on Cognitive Infocommunications (CogInfoCom). IEEE, 1–6.
Antoine Bussy, Pierre Gergondet, Abderrahmane Kheddar, François Keith, and André Crosnier. 2012. Proactive behavior of a humanoid robot in a haptic transportation task with a human partner. In 2012 IEEE RO-MAN: The 21st IEEE International Symposium on Robot and Human Interactive Communication. IEEE, 962–967.
Javier Cámara, Kirstie L. Bellman, Jeffrey O. Kephart, Marco Autili, Nelly Bencomo, Ada Diaconescu, Holger Giese, Sebastian Götz, Paola Inverardi, Samuel Kounev, and Massimo Tivoli. 2017. Self-aware computing systems: Related concepts and research areas. In Self-Aware Computing Systems. Springer, 17--49.
Tao Cao, Dayou Li, Carsten Maple, and Renxi Qiu. 2013. Human interaction based Robot Self-Learning approach for generic skill learning in domestic environment. In 2013 IEEE International Conference on Robotics and Biomimetics (ROBIO). IEEE, 203–208.
Daniel Carton, Annemarie Turnwald, Dirk Wollherr, and Martin Buss. 2013. Proactively approaching pedestrians with an autonomous mobile robot in urban environments. In Experimental Robotics. Springer, 199–214.
Joao Cartucho, Rodrigo Ventura, and Manuela Veloso. 2018. Robust object recognition through symbiotic deep learning in mobile robots. In 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2336–2341.
Sandeep P. Chinchali, Scott C. Livingston, Mo Chen, and Marco Pavone. 2019. Multi-objective optimal control for proactive decision making with temporal logic models. The International Journal of Robotics Research 38, 12--13 (2019), 1490–1512.
Henriette Cramer, Nicander Kemper, Alia Amin, and Vanessa Evers. 2009. The effects of robot touch and proactive behaviour on perceptions of human-robot interactions. In 2009 4th ACM/IEEE International Conference on Human-Robot Interaction (HRI). IEEE, 275–276.
Henriette Cramer, Nicander Kemper, Alia Amin, Bob Wielinga, and Vanessa Evers. 2009. ‘Give me a hug’: the effects of touch and autonomy on people’s responses to embodied social agents. Computer Animation and Virtual Worlds 20, 2--3 (2009), 437–445.
Kerstin Dautenhahn. 2007. Socially intelligent robots: dimensions of human–robot interaction. Philosophical Transactions of the Royal Society B: Biological Sciences 362, 1480 (2007), 679–704.
Michiel De Jong, Kevin Zhang, Aaron M. Roth, Travers Rhodes, Robin Schmucker, Chenghui Zhou, Sofia Ferreira, João Cartucho, and Manuela Veloso. 2018. Towards a robust interactive and learning social robot. In Proceedings of the 17th International Conference on Autonomous Agents and MultiAgent Systems. International Foundation for Autonomous Agents and Multiagent Systems, 883–891.
Agostino De Santis, Bruno Siciliano, Alessandro De Luca, and Antonio Bicchi. 2008. An atlas of physical human–robot interaction. Mechanism and Machine Theory 43, 3 (2008), 253–270.
Inbal Deutsch, Hadas Erel, Michal Paz, Guy Hoffman, and Oren Zuckerman. 2019. Home robotic devices for older adults: Opportunities and concerns. Computers in Human Behavior 98 (2019), 122–133.
Sandra Devin, Aurélie Clodic, and Rachid Alami. 2017. About decisions during human-robot shared plan achievement: Who should act and how?. In International Conference on Social Robotics. Springer, 453–463.
Andreea Dobra. 2014. General classification of robots. Size criteria. In 2014 23rd International Conference on Robotics in Alpe-Adria-Danube Region (RAAD). IEEE, 1–6.
Jianhao Du, Weihua Sheng, and Meiqin Liu. 2018. A human–robot collaborative system for robust three-dimensional mapping. IEEE/ASME Transactions on Mechatronics 23, 5 (2018), 2358–2368.
Eleni Efthimiou, Stavroula-Evita Fotinea, Theodore Goulas, Athanasia-Lida Dimou, Maria Koutsombogera, Vassilis Pitsikalis, Petros Maragos, and Costas Tzafestas. 2016. The MOBOT platform–showcasing multimodality in human-assistive robot interaction. In International Conference on Universal Access in Human-Computer Interaction. Springer, 382–391.
Eleni Efthimiou, Stavroula-Evita Fotinea, Theodore Goulas, Maria Koutsombogera, Panagiotis Karioris, Anna Vacalopoulou, Isidoros Rodomagoulakis, Petros Maragos, Costas Tzafestas, Vassilis Pitsikalis, Yiannis Koumpouros, Alexandra Karavasili, Panagiotis Siavelis, Foteini Koureta, and Despoina Alexopoulou. 2016. The MOBOT rollator human-robot interaction model and user evaluation process. In IEEE Symposium Series on Computational Intelligence (SSCI'16). IEEE, 1--8.
Natascha Esau and Lisa Kleinjohann. 2011. Emotional robot competence and its use in robot behavior control. In Emotional Engineering. Springer, 119–142.
David Feil-Seifer and Maja J. Matarić. 2009. Toward socially assistive robotics for augmenting interventions for children with autism spectrum disorders. In Experimental Robotics. Springer, 201–210.
Yuan Feng, Emilia I. Barakova, Suihuai Yu, Jun Hu, and G. W. Rauterberg. 2020. Effects of the level of interactivity of a social robot and the response of the augmented reality display in contextual interactions of people with dementia. Sensors 20, 13 (2020), 3771.
Alessandro Filippeschi, Lorenzo Peppoloni, Ioannis Kostavelis, Justyna Gerłowska, Emanuele Ruffaldi, Dimitris Giakoumis, Dimitrios Tzovaras, Konrad Rejdak, and Carlo Alberto Avizzano. 2018. Towards Skills Evaluation of Elderly for Human-Robot Interaction. In 2018 27th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN). IEEE, 886–892.
Julia Fink, Séverin Lemaignan, Pierre Dillenbourg, Philippe Rétornaz, Florian Vaussard, Alain Berthoud, Francesco Mondada, Florian Wille, and Karmen Franinović. 2014. Which robot behavior can motivate children to tidy up their toys? Design and Evaluation of” Ranger”. In Proceedings of the 2014 ACM/IEEE International Conference on Human-Robot Interaction. 439–446.
Stavroula-Evita Fotinea, Eleni Efthimiou, Theodore Goulas, Athanasia-Lida Dimou, Costas Tzafestas, and Vassilis Pitsikalis. 2016. The mobot human-robot interaction: Showcasing assistive hri. In Proceedings of the 20th Pan-Hellenic Conference on Informatics. Association for Computing Machinery, 1–6.
Ignasi Andrés Franch. 2017. Asking Human Help in Contingent Planning. In Proceedings of the 16th Conference on Autonomous Agents and Multiagent Systems. International Foundation for Autonomous Agents and Multiagent Systems, 1832–1833.
Dorothée François, Stuart Powell, and Kerstin Dautenhahn. 2009. A long-term study of children with autism playing with a robotic pet: Taking inspirations from non-directive play therapy to encourage children’s proactivity and initiative-taking. Interaction Studies 10, 3 (2009), 324–373.
Anais Garrell, Michael Villamizar, Francesc Moreno-Noguer, and Alberto Sanfeliu. 2013. Proactive behavior of an autonomous mobile robot for human-assisted learning. In 2013 IEEE RO-MAN. IEEE, 107–113.
Anais Garrell, Michael Villamizar, Francesc Moreno-Noguer, and Alberto Sanfeliu. 2017. Teaching robot’s proactive behavior using human assistance. International Journal of Social Robotics 9, 2 (2017), 231–249.
Dimitrios Giakoumis, Georgia Peleka, Manolis Vasileiadis, Ioannis Kostavelis, and Dimitrios Tzovaras. 2020. Service Robot Behaviour Adaptation Based on User Mood, Towards Better Personalized Support of MCI Patients at Home. In Smart Assisted Living. Springer, 209–226.
YoungCheol Go, Gee Pyeong Han, Kyo IL Chung, and Myung-Ae Chung. 2012. How to provide service using intention and service models. In 2012 14th International Conference on Advanced Communication Technology (ICACT). IEEE, 246–250.
Barbara Gonsior, Christian Landsiedel, Nicole Mirnig, Stefan Sosnowski, Ewald Strasser, Jakub Zlotowski, Martin Buss, Kolja Kühnlenz, Manfred Tscheligi, Astrid Weiss, and Dirk Wollherr. 2012. Impacts of Multimodal Feedback on Efficiency of Proactive Information Retrieval from Task-Related HRI. JACIII 16, 2 (2012), 313–326.
Barbara Gonsior, Stefan Sosnowski, Malte Buß, Dirk Wollherr, and Kolja Kühnlenz. 2012. An emotional adaption approach to increase helpfulness towards a robot. In 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems. IEEE, 2429–2436.
Enrique González, John Páez, Fernando Luis-Ferreira, João Sarraipa, and Ricardo Gonçalves. 2019. Human-robot scaffolding, an architecture to support the learning process. In Iberian Robotics Conference. Springer, 528–541.
O. Can Görür and Aydan M. Erkmen. 2019. Intention and Body-Mood Engineering via Proactive Robot Moves in HRI. In Rapid Automation: Concepts, Methodologies, Tools, and Applications. IGI Global, 247–275.
Jürgen Graf, Stephan Puls, and Heinz Wörn. 2009. Incorporating novel path planning method into cognitive vision system for safe human-robot interaction. In 2009 Computation World: Future Computing, Service Computation, Cognitive, Adaptive, Content, Patterns. IEEE, 443–447.
Jesse Gray and Cynthia Breazeal. 2014. Manipulating mental states through physical action. International Journal of Social Robotics 6, 3 (2014), 315–327.
Anders Green. 2007. Characterising dimensions of use for designing adaptive dialogues for human-robot communication. In RO-MAN 2007-The 16th IEEE International Symposium on Robot and Human Interactive Communication. IEEE, 1078–1083.
Siva Leela Krishna Chand Gudi, Suman Ojha, Sidra, Benjamin Johnston, and Mary-Anne Williams. 2017. A Proactive Robot Tutor Based on Emotional Intelligence. In International Conference on Robot Intelligence Technology and Applications. Springer, 113–120.
Xingzhi Guo, Yu-Cian Huang, Edwinn Gamborino, Shih-Huan Tseng, Li-Chen Fu, and Su-Ling Yeh. 2019. Inferring Human Feelings and Desires for Human-Robot Trust Promotion. In International Conference on Human-Computer Interaction. Springer, 365–375.
Thorsten Hempel and Ayoub Al-Hamadi. 2020. SLAM-Based Multistate Tracking System for Mobile Human-Robot Interaction. In International Conference on Image Analysis and Recognition. Springer, 368–376.
Andreas Hermann, Felix Mauch, Klaus Fischnaller, Sebastian Klemm, Arne Roennau, and Ruediger Dillmann. 2015. Anticipate your surroundings: Predictive collision detection between dynamic obstacles and planned robot trajectories on the GPU. In 2015 European Conference on Mobile Robots (ECMR). IEEE, 1–8.
Koen Hindriks, Mark A. Neerincx, and Mirek Vink. 2011. The icat as a natural interaction partner. In International Conference on Autonomous Agents and Multiagent Systems. Springer, 212–231.
Khoi Hoang Dinh, Ozgur S Oguz, Mariam Elsayed, and Dirk Wollherr. 2019. Adaptation and transfer of robot motion policies for close proximity human-robot interaction. Frontiers in Robotics and AI (2019), 69.
Philipp Hock, Johannes Kraus, Marcel Walch, Nina Lang, and Martin Baumann. 2016. Elaborating feedback strategies for maintaining automation in highly automated driving. In Proceedings of the 8th International Conference on Automotive User Interfaces and Interactive Vehicular Applications. Association for Computing Machinery, 105–112.
Mohammed Moshiul Hoque, Dipankar Das, Yoshinori Kobabashi, Yoshinori Kuno, and Kaushik Deb. 2014. Proactive approach of making eye contact with the target human in multi-party settings. In 16th Int’l Conf. Computer and Information Technology. IEEE, 191–196.
Chien-Ming Huang and Bilge Mutlu. 2016. Anticipatory robot control for efficient human-robot collaboration. In 2016 11th ACM/IEEE International Conference on Human-Robot Interaction (HRI). IEEE, 83–90.
Rui Huang, Hong Cheng, Hongliang Guo, Xichuan Lin, Qiming Chen, and Fuchun Sun. 2016. Learning cooperative primitives with physical human-robot interaction for a HUman-Powered Lower EXoskeleton. In 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 5355–5360.
Rui Huang, Hong Cheng, Jing Qiu, and Jianwei Zhang. 2019. Learning physical human–robot interaction with coupled cooperative primitives for a lower exoskeleton. IEEE Transactions on Automation Science and Engineering 16, 4 (2019), 1566–1574.
Yanrong Jiang, Nanfeng Xiao, and Jianhua Han. 2013. Automatic Control of Contextual Interaction Integrated with Affection and Architectural Attentional Control. International Journal of Advanced Robotic Systems 10, 3 (2013), 164.
Wanting Jin, Paolo Salaris, and Philippe Martinet. 2020. Proactive-Cooperative Navigation in Human-Like Environment for Autonomous Robots. In International Conference on Informatics in Control, Automation and Robotics.
William K. Juel, Frederik Haarslev, Eduardo R. Ramirez, Emanuela Marchetti, Kerstin Fischer, Danish Shaikh, Poramate Manoonpong, Christian Hauch, Leon Bodenhagen, and Norbert Krüger. 2020. Smooth robot: Design for a novel modular welfare robot. Journal of Intelligent & Robotic Systems 98, 1 (2020), 19–37.
Takayuki Kanda, Dylan F. Glas, Masahiro Shiomi, and Norihiro Hagita. 2009. Abstracting people’s trajectories for social robots to proactively approach customers. IEEE Transactions on Robotics 25, 6 (2009), 1382–1396.
Yusuke Kato, Takayuki Kanda, and Hiroshi Ishiguro. 2015. May i help you?-design of human-like polite approaching behavior. In 2015 10th ACM/IEEE International Conference on Human-Robot Interaction (HRI). IEEE, 35–42.
Harmish Khambhaita and Rachid Alami. 2017. A human-robot cooperative navigation planner. In Proceedings of the Companion of the 2017 ACM/IEEE International Conference on Human-Robot Interaction. 161–162.
Harmish Khambhaita and Rachid Alami. 2020. Viewing robot navigation in human environment as a cooperative activity. In Robotics Research. Springer, 285–300.
Mahdi Khoramshahi and Aude Billard. 2020. A dynamical system approach for detection and reaction to human guidance in physical human–robot interaction. Autonomous Robots 44, 8 (2020), 1411–1429.
Ioannis Kostavelis, Dimitrios Giakoumis, Sotiris Malasiotis, and Dimitrios Tzovaras. 2015. RAMCIP: towards a robotic assistant to support elderly with mild cognitive impairments at home. In International Symposium on Pervasive Computing Paradigms for Mental Health. Springer, 186–195.
Matthias Kraus, Nicolas Wagner, and Wolfgang Minker. 2020. Effects of Proactive Dialogue Strategies on Human-Computer Trust. In Proceedings of the 28th ACM Conference on User Modeling, Adaptation and Personalization. 107–116.
Barbara Kühnlenz, Stefan Sosnowski, Malte Buß, Dirk Wollherr, Kolja Kühnlenz, and Martin Buss. 2013. Increasing helpfulness towards a robot by emotional adaption to the user. International Journal of Social Robotics 5, 4 (2013), 457–476.
I-Han Kuo, Chandimal Jayawardena, Elizabeth Broadbent, and Bruce A. MacDonald. 2011. Multidisciplinary design approach for implementation of interactive services. International Journal of Social Robotics 3, 4 (2011), 443–456.
Woo Young Kwon and Il Hong Suh. 2010. Smart Action Selection Architecture Taking into account both Goal-orientedness and Proactivity. In 2nd International Symposium on New Frontiers in Human-Robot Interaction-A Symposium at the AISB 2010 Convention. 58–63.
Woo Young Kwon and Il Hong Suh. 2011. Towards proactive assistant robots for human assembly tasks. In Proceedings of the 6th international conference on Human-robot interaction. 175–176.
Woo Young Kwon and Il Hong Suh. 2012. A temporal bayesian network with application to design of a proactive robotic assistant. In 2012 IEEE International Conference on Robotics and Automation. IEEE, 3685–3690.
Woo Young Kwon and Il Hong Suh. 2013. Proactive planning using a hybrid temporal influence diagram for human assistive robots. In 2013 IEEE International Conference on Robotics and Automation. IEEE, 1785–1791.
Woo Young Kwon and Il Hong Suh. 2014. Planning of proactive behaviors for human–robot cooperative tasks under uncertainty. Knowledge-Based Systems 72 (2014), 81–95.
Daniel A. Lazar, Ramtin Pedarsani, Kabir Chandrasekher, and Dorsa Sadigh. 2018. Maximizing road capacity using cars that influence people. In 2018 IEEE Conference on Decision and Control (CDC). IEEE, 1801–1808.
WonHyong Lee, Chung Hyuk Park, Seyun Jang, and Hye-Kyung Cho. 2020. Design of Effective Robotic Gaze-Based Social Cueing for Users in Task-Oriented Situations: How to Overcome In-Attentional Blindness? Applied Sciences 10, 16 (2020), 5413.
Youngho Lee, Young J Ryoo, Jongmyung Choi, and Sungtae Moon. 2013. Intelligent machine space for interacting with human in Ubiquitous Virtual Reality. In International Conference on Distributed, Ambient, and Pervasive Interactions. Springer, 459–465.
Iolanda Leite, André Pereira, and Jill Fain Lehman. 2017. Persistent memory in repeated child-robot conversations. In Proceedings of the 2017 Conference on Interaction Design and Children. Association for Computing Machinery, 238–247.
Karen Leung, Edward Schmerling, Mengxuan Zhang, Mo Chen, John Talbot, J. Christian Gerdes, and Marco Pavone. 2020. On infusing reachability-based safety assurance within planning frameworks for human–robot vehicle interactions. The International Journal of Robotics Research 39, 10--11 (2020), 1326–1345.
Songpo Li and Xiaoli Zhang. 2017. Implicit intention communication in human–robot interaction through visual behavior studies. IEEE Transactions on Human-Machine Systems 47, 4 (2017), 437–448.
Zhe Li, Jannik Fritsch, Sven Wachsmuth, and Gerhard Sagerer. 2006. An object-oriented approach using a top-down and bottom-up process for manipulative action recognition. In Joint Pattern Recognition Symposium. Springer, 212–221.
Phoebe Liu, Dylan F. Glas, Takayuki Kanda, and Hiroshi Ishiguro. 2017. Two demonstrators are better than one—A social robot that learns to imitate people with different interaction styles. IEEE Transactions on Cognitive and Developmental Systems 11, 3 (2017), 319–333.
Rui Liu and Xiaoli Zhang. 2016. Fuzzy context-specific intention inference for robotic caregiving. International Journal of Advanced Robotic Systems 13, 5 (2016), 1729881416662780.
Rui Liu, Xiaoli Zhang, and Songpo Li. 2014. Use context to understand user’s implicit intentions in Activities of Daily Living. In 2014 IEEE International Conference on Mechatronics and Automation. IEEE, 1214–1219.
M. Losch, Sven Schmidt-Rohr, Steffen Knoop, Stefan Vacek, and Rüdiger Dillmann. 2007. Feature set selection and optimal classifier for human activity recognition. In RO-MAN 2007-The 16th IEEE International Symposium on Robot and Human Interactive Communication. IEEE, 1022–1027.
Michal Luria, Rebecca Zheng, Bennett Huffman, Shuangni Huang, John Zimmerman, and Jodi Forlizzi. 2020. Social Boundaries for Personal Agents in the Interpersonal Space of the Home. In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems. 1–12.
Guilherme Maeda, Aayush Maloo, Marco Ewerton, Rudolf Lioutikov, and Jan Peters. 2016. Anticipative Interaction Primitives for Human-Robot Collaboration. In AAAI Fall Symposia.
Jim Mainprice and Dmitry Berenson. 2013. Human-robot collaborative manipulation planning using early prediction of human motion. In 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems. IEEE, 299–306.
Harsh Maithani, Juan Antonio Corrales Ramon, and Youcef Mezouar. 2019. Predicting human intent for cooperative physical human-robot interaction tasks. In 2019 IEEE 15th International Conference on Control and Automation (ICCA). IEEE, 1523–1528.
Patrizia Marti, Leonardo Giusti, Alessandro Pollini, and Alessia Rullo. 2005. Experiencing the flow: design issues in human-robot interaction. In Proceedings of the 2005 Joint Conference on Smart Objects and Ambient Intelligence: Innovative Context-Aware Services: Usages and Technologies. Association for Computing Machinery, 69–74.
Martin Mason and Manuel Lopes. 2011. Robot self-initiative and personalization by learning through repeated interactions. In 2011 6th ACM/IEEE International Conference on Human-Robot Interaction (HRI). IEEE, 433–440.
Elias Matsas, George-Christopher Vosniakos, and Dimitris Batras. 2018. Prototyping proactive and adaptive techniques for human-robot collaboration in manufacturing using virtual reality. Robotics and Computer-Integrated Manufacturing 50 (2018), 168–180.
José Ramón Medina, Tamara Lorenz, Dongheui Lee, and Sandra Hirche. 2012. Disagreement-aware physical assistance through risk-sensitive optimal feedback control. In 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems. IEEE, 3639–3645.
S. S. Mehta, E. A. Doucette, and J. W. Curtis. 2017. Curious partner: an autonomous system that proactively dialogues with human teammates. In Micro-and Nanotechnology Sensors, Systems, and Applications IX, Vol. 10194. International Society for Optics and Photonics, 101941W.
Pierrick Milhorat, Divesh Lala, Koji Inoue, Tianyu Zhao, Masanari Ishida, Katsuya Takanashi, Shizuka Nakamura, and Tatsuya Kawahara. 2019. A conversational dialogue manager for the humanoid robot ERICA. In Advanced Social Interaction with Agents. Springer, 119–131.
Gregoire Milliez. 2018. Buddy: A companion robot for the whole family. In Companion of the 2018 ACM/IEEE International Conference on Human-Robot Interaction. 40–40.
Brian Mok. 2016. Effects of proactivity and expressivity on collaboration with interactive robotic drawers. In 2016 11th ACM/IEEE International Conference on Human-Robot Interaction (HRI). IEEE, 633–634.
Brian Mok, Stephen Yang, David Sirkin, and Wendy Ju. 2015. Collaboration with Robotic Drawers. In Proceedings of the Tenth Annual ACM/IEEE International Conference on Human-Robot Interaction Extended Abstracts. Association for Computing Machinery, 291–291.
Brian Ka-Jun Mok, Stephen Yang, David Sirkin, and Wendy Ju. 2015. A place for every tool and every tool in its place: Performing collaborative tasks with interactive robotic drawers. In 2015 24th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN). IEEE, 700–706.
Christophe Mollaret, Alhayat Ali Mekonnen, Isabelle Ferrané, Julien Pinquier, and Frédéric Lerasle. 2015. Perceiving user’s intention-for-interaction: A probabilistic multimodal data fusion scheme. In 2015 IEEE International Conference on Multimedia and Expo (ICME). IEEE, 1–6.
Clément Moulin-Frier, Tobias Fischer, Maxime Petit, Grégoire Pointeau, Jordi-Ysard Puigbo, Ugo Pattacini, Sock Ching Low, Daniel Camilleri, Phuong Nguyen, Matej Hoffmann, Hyung Jin Chang, Martina Zambelli, Anne-Laure Mealier, Andreas Damianou, Giorgio Metta, Tony J. Prescott, Yiannis Demiris, Peter Ford Dominey, and Paul F. M. J. Verschure. 2017. DAC-h3: a proactive robot cognitive architecture to acquire and express knowledge about the world and the self. IEEE Transactions on Cognitive and Developmental Systems 10, 4 (2017), 1005--1022.
E. Mumolo, M. Nolich, and G. Vercelli. 2001. Pro-active service robots in a health care framework: vocal interaction using natural language and prosody. In Proceedings 10th IEEE International Workshop on Robot and Human Interactive Communication. ROMAN 2001 (Cat. No. 01TH8591). IEEE, 606–611.
Junya Nakanishi, Itaru Kuramoto, Jun Baba, Kohei Ogawa, Yuichiro Yoshikawa, and Hiroshi Ishiguro. 2019. Soliloquising social robot in a hotel room. In Proceedings of the 31st Australian Conference on Human-Computer-Interaction. Association for Computing Machinery, 21–29.
Mihai Nan, Alexandra Stefania Ghiță, Alexandru-Florin Gavril, Mihai Trascau, Alexandru Sorici, Bogdan Cramariuc, and Adina Magda Florea. 2019. Human action recognition for social robots. In 2019 22nd International Conference on Control Systems and Computer Science (CSCS). IEEE, 675–681.
Heramb Nemlekar, Dharini Dutia, and Zhi Li. 2019. Object transfer point estimation for fluent human-robot handovers. In 2019 International Conference on Robotics and Automation (ICRA). IEEE, 2627–2633.
Victor Ng-Thow-Hing, Jongwoo Lim, Joel Wormer, Ravi Kiran Sarvadevabhatla, Carlos Rocha, Kikuo Fujimura, and Yoshiaki Sakagami. 2008. The memory game: Creating a human-robot interactive scenario for asimo. In 2008 IEEE/RSJ International Conference on Intelligent Robots and Systems. IEEE, 779–786.
Davide Nicolis, Andrea Maria Zanchettin, and Paolo Rocco. 2018. Human intention estimation based on neural networks for enhanced collaboration with robots. In 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 1326–1333.
Dimitri Ognibene, Lorenzo Mirante, and Letizia Marchegiani. 2019. Proactive Intention Recognition for Joint Human-Robot Search and Rescue Missions through Monte-Carlo Planning in POMDP Environments. In International Conference on Social Robotics. Springer, 332–343.
Ozgur S. Oguz, Volker Gabler, Gerold Huber, Zhehua Zhou, and Dirk Wollherr. 2016. Hybrid human motion prediction for action selection within human-robot collaboration. In International Symposium on Experimental Robotics. Springer, 289–298.
Ozgur S. Oguz, Omer C. Sari, Khoi H. Dinh, and Dirk Wollherr. 2017. Progressive stochastic motion planning for human-robot interaction. In 2017 26th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN). IEEE, 1194–1201.
Joao Lobato Oliveira, Gökhan Ince, Keisuke Nakamura, Kazuhiro Nakadai, Hiroshi G. Okuno, Luis Paulo Reis, and Fabien Gouyon. 2012. An active audition framework for auditory-driven HRI: Application to interactive robot dancing. In 2012 IEEE RO-MAN: The 21st IEEE International Symposium on Robot and Human Interactive Communication. IEEE, 1078–1085.
Amit Kumar Pandey, Muhammad Ali, and Rachid Alami. 2013. Towards a task-aware proactive sociable robot based on multi-state perspective-taking. International Journal of Social Robotics 5, 2 (2013), 215–236.
Amit Kumar Pandey, Muhammad Ali, Matthieu Warnier, and Rachid Alami. 2011. Towards multi-state visuo-spatial reasoning based proactive human-robot interaction. In 2011 15th International Conference on Advanced Robotics (ICAR). IEEE, 143–149.
Kwang-Hyun Park, Hyong-Euk Lee, Youngmin Kim, and Z. Zenn Bien. 2008. A steward robot for human-friendly human-machine interaction in a smart house environment. IEEE Transactions on Automation Science and Engineering 5, 1 (2008), 21–25.
Georgia Peleka, Andreas Kargakos, Evangelos Skartados, Ioannis Kostavelis, Dimitrios Giakoumis, Iason Sarantopoulos, Zoe Doulgeri, Michalis Foukarakis, Margherita Antona, Sandra Hirche, Emannuele Ruffaldi, Bartlomiej Stanczyk, Anastasios Zompas, Joan Hernandez-Farigola, Natalia Roberto, Konrad Rejdak, and Dimitrios Tzovaras. 2018. RAMCIP-A service robot for MCI patients at home. In 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 1–9.
Zhenhui Peng, Yunhwan Kwon, Jiaan Lu, Ziming Wu, and Xiaojuan Ma. 2019. Design and evaluation of service robot’s proactivity in decision-making support process. In Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems. Association for Computing Machinery, 1–13.
Michele Persiani and Thomas Hellström. 2020. Intent Recognition from Speech and Plan Recognition. In International Conference on Practical Applications of Agents and Multi-Agent Systems. Springer, 212–223.
Giovanni Pezzulo, Gianluca Baldassarre, Amedeo Cesta, and Stefano Nolfi. 2011. Research on cognitive robotics at the institute of cognitive sciences and technologies, national research council of italy. Cognitive Processing 12, 4 (2011), 367–374.
Manuel Pinheiro and Estela Bicho. 2013. A socially assistive robot for people with motor impairments. In 2013 IEEE 3rd Portuguese Meeting in Bioengineering (ENBENG). IEEE, 1–7.
Md Golam Rashed, Royta Suzuki, Toshiki Kikugawa, Antony Lam, Yoshinori Kobayashi, and Yoshinori Kuno. 2015. Network guide robot system proactively initiating interaction with humans based on their local and global behaviors. In International Conference on Intelligent Computing. Springer, 283–294.
M. Golam Rashed, Ryota Suzuki, Antony Lam, Yoshinori Kobayashi, and Yoshinori Kuno. 2015. Toward museum guide robots proactively initiating interaction with humans. In Proceedings of the Tenth Annual ACM/IEEE International Conference on Human-Robot Interaction Extended Abstracts. Association for Computing Machinery, 1–2.
M. Golam Rashed, R. Suzuki, A. Lam, Y. Kobayashi, and Y. Kuno. 2015. A vision based guide robot system: Initiating proactive social human robot interaction in museum scenarios. In 2015 International Conference on Computer and Information Engineering (ICCIE). IEEE, 5–8.
Claire Rivoire and Angelica Lim. 2016. Habit detection within a long-term interaction with a social robot: an exploratory study. In Proceedings of the International Workshop on Social Learning and Multimodal Interaction for Designing Artificial Agents. Association for Computing Machinery, 1–6.
Ben Robins, Kerstin Dautenhahn, Rene Te Boekhorst, and Aude Billard. 2005. Robotic assistants in therapy and education of children with autism: can a small humanoid robot help encourage social interaction skills? Universal Access in the Information Society 4, 2 (2005), 105–120.
Stephanie Rosenthal, Manuela Veloso, and Anind K. Dey. 2012. Is someone in this office available to help me? Journal of Intelligent & Robotic Systems 66, 1 (2012), 205--221.
Behzad Sadrfaridpour, Maziar Fooladi Mahani, Zhanrui Liao, and Yue Wang. 2018. Trust-Based Impedance Control Strategy for Human-Robot Cooperative Manipulation. In ASME 2018 Dynamic Systems and Control Conference. American Society of Mechanical Engineers Digital Collection.
Audun Sanderud, Trygve Thomessen, Hisashi Osumi, and Mihoko Niitsuma. 2015. A proactive strategy for safe human-robot collaboration based on a simplified risk analysis. Modeling, Identification and Control 36, 1 (2015), 11–21. DOI:
Audun Rønning Sanderud, Mihoko Niitsuma, and Trygve Thomessen. 2015. A likelihood analysis for a risk analysis for safe human robot collaboration. In 2015 IEEE 20th Conference on Emerging Technologies & Factory Automation (ETFA). IEEE, 1–6.
Andreas J. Schmid, Oliver Weede, and Heinz Worn. 2007. Proactive robot task selection given a human intention estimate. In RO-MAN 2007-The 16th IEEE International Symposium on Robot and Human Interactive Communication. IEEE, 726–731.
Andreas J. Schmid, Heinz Wörn, Oliver C. Schrempf, and Uwe D. Hanebeck. 2006. Towards intuitive human-robot cooperation. In Proceedings of the 2nd International Workshop on Human-Centered Robotic Systems (HCRS). 7–12.
Sven R. Schmidt-Rohr, Martin Loesch, Zhixing Xue, and Rüdiger Dillmann. 2008. Hardware and Software Architecture for Robust Autonomous Behavior of a Domestic Robot Assistant. In Intelligent Autonomous Systems 10. IOS Press, 74–83.
Oliver C. Schrempf, Uwe D. Hanebeck, Andreas J. Schmid, and Heinz Worn. 2005. A novel approach to proactive human-robot cooperation. In ROMAN 2005. IEEE International Workshop on Robot and Human Interactive Communication, 2005. IEEE, 555–560.
Ruth Schulz, Philipp Kratzer, and Marc Toussaint. 2017. Building a bridge with a robot: a system for collaborative on-table task execution. In Proceedings of the 5th International Conference on Human Agent Interaction. 399–403.
Alessandra Sciutti, Ambra Bisio, Francesco Nori, Giorgio Metta, Luciano Fadiga, Thierry Pozzo, and Giulio Sandini. 2012. Measuring human-robot interaction through motor resonance. International Journal of Social Robotics 4, 3 (2012), 223–234.
Alessandra Sciutti, Ambra Bisio, Francesco Nori, Giorgio Metta, Luciano Fadiga, and Giulio Sandini. 2013. Robots can be perceived as goal-oriented agents. Interaction Studies 14, 3 (2013), 329–350.
Lanbo She and Joyce Chai. 2017. Interactive learning of grounded verb semantics towards human-robot communication. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Association for Computational Linguistics, 1634–1644.
Weihua Sheng, Anand Thobbi, and Ye Gu. 2014. An integrated framework for human–robot collaborative manipulation. IEEE Transactions on Cybernetics 45, 10 (2014), 2030–2041.
Chao Shi, Masahiro Shiomi, Christian Smith, Takayuki Kanda, and Hiroshi Ishiguro. 2013. A model of handing interaction towards a pedestrian. In 2013 8th ACM/IEEE International Conference on Human-Robot Interaction (HRI). IEEE, 415–415.
Dadhichi Shukla, Özgür Erkent, and Justus Piater. 2017. Proactive, incremental learning of gesture-action associations for human-robot collaboration. In 2017 26th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN). IEEE, 346–353.
Rui Silva, Francisco S. Melo, and Manuela Veloso. 2016. Adaptive symbiotic collaboration for targeted complex manipulation tasks. In Proceedings of the Twenty-second European Conference on Artificial Intelligence. IOS Press, 863–870.
Chapa Sirithunge, A. G. Buddhika, P. Jayasekara, and D. P. Chandima. 2019. Proactive robots with the perception of nonverbal human behavior: A review. IEEE Access 7 (2019), 77308–77327.
Chapa Sirithunge, H. M. Ravindu, T. Bandara, A. G. Buddhika, P. Jayasekara, and D. P. Chandima. 2019. Situation Awareness for Proactive Robots in HRI. In 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 7813–7820.
Lenja Sorokin, Ronee Chadowitz, and Nina Kauffmann. 2019. A change of perspective: Designing the automated vehicle as a new social actor in a public space. In Extended Abstracts of the 2019 CHI Conference on Human Factors in Computing Systems. Association for Computing Machinery, 1–8.
Aaron Steinfeld, Terrence Fong, David Kaber, Michael Lewis, Jean Scholtz, Alan Schultz, and Michael Goodrich. 2006. Common metrics for human-robot interaction. In Proceedings of the 1st ACM SIGCHI/SIGART Conference on Human-Robot Interaction. 33–40.
Young-Ho Suh, Hyun Kim, Joo-Haeng Lee, Joonmyun Cho, Moohun Lee, Jeongnam Yeom, and Eun-Sun Cho. 2011. Future robotic computer: A new type of computing device with robotic functions. In Proceedings of the 6th International Conference on Human-Robot Interaction. Association for Computing Machinery, 261–262.
Kaoru Sumi. 2012. Human Interface of Robots or Agents via Facial and Word Expression. In International Symposium on Artificial Life and Robotics (AROB). ISAROB.
Indranil Sur and Heni Ben Amor. 2017. Robots that anticipate pain: Anticipating physical perturbations from visual cues through deep predictive models. In 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 5541–5548.
Hao Tan, Ying Zhao, Shiyan Li, Wei Wang, Ming Zhu, Jie Hong, and Xiang Yuan. 2020. Relationship between social robot proactive behavior and the human perception of anthropomorphic attributes. Advanced Robotics 34, 20 (2020), 1324–1336.
Kuo-Hao Tang, Chia-Feng Ho, Jan Mehlich, and Shih-Ting Chen. 2020. Assessment of Handover Prediction Models in Estimation of Cycle Times for Manual Assembly Tasks in a Human–Robot Collaborative Environment. Applied Sciences 10, 2 (2020), 556.
Priyesh Tiwari, Jim Warren, Karen Day, Bruce MacDonald, Chandimal Jayawardena, I. Han Kuo, Aleksandar Igic, and Chandan Datta. 2011. Feasibility study of a robotic medication assistant for the elderly. In Proceedings of the Twelfth Australasian User Interface Conference-Volume 117. Australian Computer Society, 57–66.
Elena Torta, Franz Werner, David O. Johnson, James F. Juola, Raymond H. Cuijpers, Marco Bazzani, Johannes Oberzaucher, John Lemberger, Hadas Lewy, and Joseph Bregman. 2014. Evaluation of a small socially-assistive humanoid robot in intelligent homes for the care of the elderly. Journal of Intelligent & Robotic Systems 76, 1 (2014), 57–71.
Xuan-Tung Truong and Trung Dung Ngo. 2017. Toward socially aware robot navigation in dynamic and crowded environments: A proactive social motion model. IEEE Transactions on Automation Science and Engineering 14, 4 (2017), 1743–1760.
Xuan-Tung Truong, Voo Nyuk Yoong, and Dung Ngo Trung. 2016. Approaching humans in crowded and dynamic environments. In 2016 IEEE International Conference on Advanced Intelligent Mechatronics (AIM). IEEE, 476–481.
Panagiota Tsarouchi, Sotiris Makris, and George Chryssolouris. 2016. Human–robot interaction review and challenges on task planning and programming. International Journal of Computer Integrated Manufacturing 29, 8 (2016), 916–931.
Vincent Wing Sun Tung and Norman Au. 2018. Exploring customer experiences with robotics in hospitality. International Journal of Contemporary Hospitality Management (2018).
Alessandro Umbrico, Amedeo Cesta, Gabriella Cortellessa, and Andrea Orlandini. 2018. A goal triggering mechanism for continuous human-robot interaction. In International Conference of the Italian Association for Artificial Intelligence. Springer, 460–473.
Gentiane Venture, Ritta Baddoura, and Tianxiang Zhang. 2013. Experiencing the familiar, understanding the interaction & responding to a robot proactive partner. In 2013 8th ACM/IEEE International Conference on Human-Robot Interaction (HRI). IEEE, 247–248.
Raquel Viciana-Abad, Rebeca Marfil, Jose M. Perez-Lorenzo, Juan P. Bandera, Adrian Romero-Garces, and Pedro Reche-Lopez. 2014. Audio-visual perception system for a humanoid robotic head. Sensors 14, 6 (2014), 9522–9545.
Manuel Vinagre, Joan Aranda, and Alicia Casals. 2014. An interactive robotic system for human assistance in domestic environments. In International Conference on Computers for Handicapped Persons. Springer, 152–155.
Zhikun Wang, Katharina Mülling, Marc Peter Deisenroth, Heni Ben Amor, David Vogt, Bernhard Schölkopf, and Jan Peters. 2013. Probabilistic movement modeling for intention inference in human–robot interaction. The International Journal of Robotics Research 32, 7 (2013), 841–858.
Kenton Williams, José Acevedo Flores, and Joshua Peters. 2014. Affective robot influence on driver adherence to safety, cognitive load reduction and sociability. In Proceedings of the 6th International Conference on Automotive User Interfaces and Interactive Vehicular Applications. Association for Computing Machinery, 1–8.
Weronika Wojtak, Flora Ferreira, Paulo Vicente, Luis Louro, Estela Bicho, and Wolfram Erlhagen. 2020. A neural integrator model for planning and value-based decision making of a robotics assistant. Neural Computing and Applications 33 (2020), 1--20.
Jingkang Xia, Deqing Huang, Yanan Li, and Na Qin. 2020. Iterative learning of human partner’s desired trajectory for proactive human–robot collaboration. International Journal of Intelligent Robotics and Applications 4 (2020), 229–242.
Yuichiro Yoshikawa, Takamasa Iio, Tsunehiro Arimoto, Hiroaki Sugiyama, and Hiroshi Ishiguro. 2017. Proactive Conversation between Multiple Robots to Improve the Sense of Human–Robot Conversation. In AAAI Fall Symposium - Technical Report. AI Access Foundation, 289–294.
Setareh Zafari and Sabine T. Koeszegi. 2020. Attitudes Toward Attributed Agency: Role of Perceived Control. International Journal of Social Robotics (2020), 1–10.
Shujing Zhang, Jingyan Qin, Sha Cao, and Jinhua Dou. 2018. HRI design research for intelligent household service robots: Teler as a case study. In International Conference of Design, User Experience, and Usability. Springer, 513–524.
Yu Zhang, Vignesh Narayanan, Tathagata Chakraborti, and Subbarao Kambhampati. 2015. A human factors analysis of proactive support in human-robot teaming. In 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 3586–3593.
Yu Zhang, Sarath Sreedharan, Anagha Kulkarni, Tathagata Chakraborti, Hankz Hankui Zhuo, and Subbarao Kambhampati. 2017. Plan explicability and predictability for robot task planning. In 2017 IEEE international conference on robotics and automation (ICRA). IEEE, 1313–1320.
Kuanhao Zheng, Dylan F. Glas, Takayuki Kanda, Hiroshi Ishiguro, and Norihiro Hagita. 2013. Supervisory control of multiple social robots for navigation. In 2013 8th ACM/IEEE International Conference on Human-Robot Interaction (HRI). IEEE, 17–24.
Tian Zhou and Juan Pablo Wachs. 2018. Early prediction for physical human robot collaboration in the operating room. Autonomous Robots 42, 5 (2018), 977–995.
Tian Zhou and Juan P. Wachs. 2019. Spiking Neural Networks for early prediction in human–robot collaboration. The International Journal of Robotics Research 38, 14 (2019), 1619–1643.
HRI '16: The Eleventh ACM/IEEE International Conference on Human Robot Interaction
Collaborative robots are quickly gaining momentum in real-world settings. This has motivated many new research questions in human-robot collaboration. In this paper, we address the questions of whether and when a robot should take initiative during ...
Advances in Artificial Intelligence – IBERAMIA 2012
Abstract
This paper proposes a method to develop an omnidirectional kick behavior for a humanoid robot. The objective is to provide a humanoid with the ability to kick in different directions and to make kicks look more like those of a human player. This ...
