Introducing CARESSER: A framework for in situ learning robot social assistance from expert knowledge and demonstrations

User Modeling and User-Adapted Interaction, Год журнала: 2022, Номер 33(2), С. 441 - 496

Опубликована: Март 12, 2022

Socially assistive robots have the potential to augment and enhance therapist's effectiveness in repetitive tasks such as cognitive therapies. However, their contribution has generally been limited domain experts not fully involved entire pipeline of design process well automatisation robots' behaviour. In this article, we present aCtive leARning agEnt aSsiStive bEhaviouR (CARESSER), a novel framework that actively learns robotic behaviour by leveraging expertise (knowledge-driven approach) demonstrations (data-driven approach). By exploiting hybrid approach, presented method enables situ fast learning, autonomous fashion, personalised patient-specific policies. With purpose evaluating our framework, conducted two user studies daily care centre which older adults affected mild dementia impairment (N = 22) were requested solve exercises with support therapist later on robot endowed CARESSER. Results showed that: (i) managed keep patients' performance stable during sessions even more so than therapist; (ii) assistance offered eventually matched preferences. We conclude CARESSER, its stakeholder-centric design, can pave way new AI approaches learn human-human interactions along human expertise, benefits speeding up learning process, eliminating need for complex reward functions, finally avoiding undesired states.

Язык: Английский

---

Self-Explaining Social Robots: An Explainable Behavior Generation Architecture for Human-Robot Interaction

Frontiers in Artificial Intelligence, Год журнала: 2022, Номер 5

Опубликована: Апрель 29, 2022

In recent years, the ability of intelligent systems to be understood by developers and users has received growing attention. This holds in particular for social robots, which are supposed act autonomously vicinity human known raise peculiar, often unrealistic attributions expectations. However, explainable models that, on one hand, allow a robot generate lively autonomous behavior and, other, enable it provide human-compatible explanations this missing. order develop such self-explaining robot, we have equipped with own needs that trigger intentions proactive behavior, form basis understandable self-explanations. Previous research shown undesirable is rated more positively after receiving an explanation. We thus aim equip capability automatically verbal its tracing internal decision-making routes. The goal way generally interpretable, therefore socio-behavioral level increasing users' understanding robot's behavior. article, present interaction architecture, designed set out requirements generation architectures propose socio-interactive framework human-robot interactions enables explaining elaborating according explanation emerge within interaction. Consequently, introduce interactive dialog flow concept incorporates empirically validated types. These concepts realized architecture integrated processing modules. components explain their integration behaviors as well Lastly, report results from qualitative evaluation working prototype laboratory setting, showing (1) able naturalistic (2) verbally self-explain user line requests.

Язык: Английский

---

Understanding Factors that Shape Children’s Long Term Engagement with an In-Home Learning Companion Robot

Interaction Design and Children, Год журнала: 2022, Номер unknown

Опубликована: Июнь 24, 2022

Social robots are emerging as learning companions for children, and research shows that they facilitate the development of interest even through brief interactions. However, little is known about how such technologies might support these goals in authentic environments over long-term periods use interaction. We designed a companion robot capable supporting children reading popular-science books by expressing social informational commentaries. deployed homes 14 families with aged 10–12 four weeks during summer. Our analysis revealed critical factors affected children's engagement adoption robot, including external vacations, family visits, extracurricular activities; family/parental involvement; individual interests. present in-depth cases illustrate demonstrate their impact on experiences discuss implications our findings design.

Язык: Английский

---

Designing Social Robot for Adults Using Self-Determination Theory and AI Technologies

IEEE Transactions on Learning Technologies, Год журнала: 2023, Номер 16(2), С. 206 - 218

Опубликована: Фев. 28, 2023

In comparison to children and young students, adult learners usually exhibit more complex learning behaviors psychological needs during the process. Designing social robots for has, thus, been a challenging task far less explored area, it requires great efforts from both technical theoretical perspectives. We, therefore, first propose novel framework that exploits latest artificial intelligence (AI) technologies established theory robot design. Under proposed framework, we implement deploy in context, which demands provide natural interactions autonomous supports learners. The evaluation results show significantly improves learners' intrinsic motivation, have also shown interests communicating with robot. This article sheds light on how design interactive contributes concrete solution employs theories as guidelines AI models enabling technologies.

Язык: Английский

---

Integrating Flow Theory and Adaptive Robot Roles: A Conceptual Model of Dynamic Robot Role Adaptation for the Enhanced Flow Experience in Long-term Multi-person Human-Robot Interactions

Опубликована: Март 10, 2024

In this paper, we introduce a novel conceptual model for robot's behavioral adaptation in its long-term interaction with humans, integrating dynamic robot role principles of flow experience from psychology. This conceptualization introduces hierarchical objective grounded the experience, serving as overarching goal robot. intertwines both cognitive and affective sub-objectives incorporates individual group-level human factors. The approach is cornerstone our model, highlighting ability to fluidly adapt support roles - leader follower aim maintaining equilibrium between activity challenge user skill, thereby fostering user's optimal experiences. Moreover, work delves into comprehensive exploration limitations potential applications proposed conceptualization. Our places particular emphasis on multi-person HRI paradigm, dimension that under-explored challenging. doing so, aspire extend applicability relevance within field, contributing future development adaptive social robots capable sustaining interactions humans.

Язык: Английский

---