User profiling and behavioral adaptation for HRI: A survey
Introduction
The development of Social Assistive Robotics (SAR) [1] applications challenges researchers to build and design socially intelligent robots that can collaborate with people. In such domains, Human-Robot Interaction (HRI) effectiveness has not only to rely on the skills of trained users but also on the ability of the robot to adapt to the users’ behavior and needs as well [2].
According to [3], one of the required characteristics of a socially interactive robot should be to perceive, learn, and recognize models of the other agents it is interacting with. Hence, while the correct perception of the human being and of his/her movements is a requirement to achieve the proper collaboration and interaction, such perception should lead the robot to profile the user’s preferences during a physical interaction with the robot, or, more generally, regarding the user’s physical capabilities. However, a complete model of the user should include also his/her cognitive state, in terms, for example, of the intentions behind the interaction or his/her internal state, and, more generally, his/her preferences regarding social interaction characteristics.
Moreover, to be effective, a robot should also be able to modify and adapt its behavior accordingly. Indeed, for improved and natural human-robot cooperation, human users will learn how to interact with the robot but, at the same time, the robotic systems should adapt to the users [2]. This adaptation requires learning a model of human behavior and integrating it into the robot physical movements, the robot decision-making algorithm [4], and the social interaction strategies. Furthermore, Castellano et al. [5] mention personalization, which they define as the ability to adapt to a specific user over time, as a key requirement for long-term socially interactive companions. Nahum-Shani et al. [6] also discuss the term customization and individualization and the need to adapt to the user based on static parameters (i.e., gender, personality) and on dynamic parameters (i.e., changes in psychological distress, response to an intervention or task) to better make decisions during the course of the interaction and the task.
The development of robotic systems capable of correctly modeling and recognizing the human behavior and of adapting their own behavior with respect to the user is a very critical task, especially in the domain of assistive robotics when working with vulnerable user populations [7]. Adaptability plays an important role in the Almere model of the acceptance of assistive social agent by older adults [8]. As observed by [9], while elderly users want to retain control over assistive devices, they prefer adaptive systems over fully user-configured ones.
In this survey, we introduce research topics addressed in this context by providing some pointers to the current literature on user profiling and behavioral adaptation. These two aspects open a complex search space that covers different research areas beyond HRI. To structure the search space and to cover a diversity of works, we introduce a classification scheme based on the concepts of physical, cognitive, and social interaction for both the profiling and behavioral adaptation aspects. Based on this scheme, we selected over 60 articles from the HRI literature published within the last five years. Our survey does not aim to be exhaustive, but, more importantly, to highlight that these two aspects are both fundamental in the development of effective and well accepted social robot. There is a closed loop relationship between profiling and adaptation, as the typical action-perception cycle, that should be addressed both from the robot point of view as well as from the interacting user’s one (see Fig. 1).
Section snippets
User profiling
Understanding more about users enables a robot to adapt its behavior with respect to their characteristics and preferences, hence to enhance the user satisfaction and robot acceptance. User preferences and defining profiles can be explicitly determined, by getting users’ response/feedback to information/questionnaire. Stereotyping is an example of user classification, based on the measurement of pre-defined features [10]. However, while explicit information is accurate, users refrain from
Behavioral adaptation
During an interaction, the ability to adapt its own behavior with respect to the behavior of the others is a fundamental characteristic that affects the effectiveness and the naturalness of the interaction itself [2]. If the flow of the interaction is not smooth, for example, people might even be irritated [70].
Profiling capabilities are of no use if the robot is not able to modify its own behavior accordingly. As in the previous section, here, we discuss the current literature dealing with
Conclusions
In this paper, we presented a survey of the recent literature dealing with the possibility of profiling a user during a human-robot interaction and of approaches that proposed a behavioral adaptation with respect to such profiles. The aim of this survey is, of course, not to be a comprehensive one but to try to cover a variety of approaches to highlight some of the key themes in the context of user profiling mechanism and behavioral adaptation.
We tried to classify such approaches with respect
Acknowledgment
This work has been partially supported by MIUR (Italian Ministry of Education, Universities, and Research) (Cod. 2015KBL78T) within the PRIN 2015 research project “UPA4SAR - User-centered Profiling and Adaptation for Socially Assistive Robotics”.
References (118)
- et al.
Centinela: a human activity recognition system based on acceleration and vital sign data
J. Pervasive Mob. Comput.
(2012) - et al.
A survey on fall detection: principles and approaches
Neurocomputing
(2013) - et al.
A survey of robot learning from demonstration
Rob. Auton. Syst.
(2009) Dimensions of personality: 16, 5 or 3?—criteria for a taxonomic paradigm
Pers. Individ. Dif.
(1991)- et al.
Investigating joint attention mechanisms through spoken humanrobot interaction
Cognition
(2011) - et al.
Human-aware robot navigation: a survey
Rob. Auton. Syst.
(2013) A review of verbal and non-verbal human-robot interactive communication
Robot. Auton. Syst.
(2015)- et al.
Human-aware planning for robots embedded in ambient ecologies
Pervasive Mob. Comput.
(2012) - et al.
Defining socially assistive robotics
9th International Conference on Rehabilitation Robotics
(2005) - et al.
Adapting robot behavior for human–robot interaction
IEEE Trans. Rob.
(2008)
A survey of socially interactive robots
Rob. Auton. Syst.
Efficient model learning from joint-action demonstrations for human-robot collaborative tasks
Proceedings of the Tenth Annual ACM/IEEE International Conference on Human-Robot Interaction
Long-term affect sensitive and socially interactive companions
Proceedings of the 4th International Workshop on Human-Computer Conversation
Building health behavior models to guide the development of just-in-time adaptive interventions: a pragmatic framework
Health Psychol.
User-robot personality matching and assistive robot behavior adaptation for post-stroke rehabilitation therapy
Intell. Serv. Rob.
Assessing acceptance of assistive social agent technology by older adults: the Almere model
Int. J. Soc. Rob.
How elderly users of a socially interactive robot experience adaptiveness, adaptability and user control
Computational Intelligence and Informatics (CINTI), 2011 IEEE 12th International Symposium on
Robots that stereotype: creating and using categories of people for human-robot interaction
J. Human-Robot Interact.
Handbook of Human Factors in Web Design, Second Edition
Human activity analysis: a review
ACM Comput. Surv.
An extensible architecture for robust multi modal human-robot communication
Intelligent Robots and Systems (IROS), 2013 IEEE/RSJ International Conference on
Wearable inertial sensors: applications, challenges, and public test benches
IEEE Robot. Autom. Mag.
Social activity recognition based on probabilistic merging of skeleton features with proximity priors from RGB-D data
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
Parallel detection of conversational groups of free-standing people and tracking of their lower-body orientation
Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
A survey on human activity recognition using wearable sensors
IEEE Commun. Surv. Tutorials
Unstructured human activity detection from RGB-D images
IEEE International Conference on Robotics and Automation
A probabilistic approach for human everyday activities recognition using body motion from RGB-D images
The 23rd IEEE International Symposium on Robot and Human Interactive Communication
Robot semantic mapping through wearable sensor-based human activity recognition
Robotics and Automation (ICRA), 2012 IEEE International Conference on
Towards online and personalized daily activity recognition, habit modeling, and anomaly detection for the solitary elderly through unobtrusive sensing
Multimedia Tools Appl.
Sensor-based Bayesian detection of anomalous living patterns in a home setting
Pers. Ubiquitous Comput.
Sensory Profile: User’s Manual
Individuals with autism: analysis of the first interaction with Nao robot based on their proprioceptive and kinematic profiles
Advances in Robot Design and Intelligent Control
Adolescent-Adult Sensory Profile: User’s manual
Impact of sensory preferences of individuals with autism on the recognition of emotions expressed by two robots, an avatar, and a human
Auton. Robots
Incremental kinesthetic teaching of motion primitives using the motion refinement tube
Auton Robots
Learning preferences for manipulation tasks from online coactive feedback
Int. J. Rob. Res.
User modelling for personalised dressing assistance by humanoid robots
Proceedings of the 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2015)
Prediction of intent in robotics and multi-agent systems.
Cogn. Process.
Theory of mind for a humanoid robot
Auton. Robots
Intent recognition for human-robot interaction
Interaction Challenges for Intelligent Assistants, Papers from the 2007 AAAI Spring Symposium, 26–28 March 2007, Stanford University, CA, USA
User models in dialog systems
User Models in Dialog Systems
Natural language processing and user modeling: synergies and limitations
User. Model. User-adapt. Interact.
Facilitating mental modeling in collaborative human-robot interaction through adverbial cues
Proceedings of the SIGDIAL 2011 Conference
A Bayesian approach for task recognition and future human activity prediction
The 23rd IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN)
Learning Dynamic Systems for Intention Recognition in Human-Robot-Cooperation
Proactive behavior of a humanoid robot in a haptic transportation task with a human partner
IEEE RO-MAN: The 21st IEEE International Symposium on Robot and Human Interactive Communication
Human intent prediction using Markov decision processes
J. Aerosp. Inf. Syst.
Proactive premature intention estimation for intuitive human-robot collaboration
IEEE/RSJ International Conference on Intelligent Robots and Systems
Evolving hidden Markov model based human intention learning and inference
2015 IEEE International Conference on Robotics and Biomimetics (ROBIO)
A temporal Bayesian network with application to design of a proactive robotic assistant
2012 IEEE International Conference on Robotics and Automation
Cited by (135)
A multi-task network for speaker and command recognition in industrial environments
2023, Pattern Recognition LettersSupporting Long-Term HRI Through Shared Family Routines
2024, ACM/IEEE International Conference on Human-Robot InteractionPersonalization of industrial human–robot communication through domain adaptation based on user feedback
2024, User Modeling and User-Adapted InteractionRedefining User Expectations: The Impact of Adjustable Social Autonomy in Human–Robot Interaction
2024, Electronics (Switzerland)