Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 106
Encontrada(s) 2 página(s)

Dipòsit Digital de la UB
Publication Artículo científico (JournalArticle)

The effects of gender and personality of robot assistants on customers' acceptance of their service

  • Forgas Coll, Santiago
  • Huertas García, Rubén
  • Andriella, Antonio
  • Alenyà, Guillem
The Covid-19 pandemic has stimulated the use of social robots in front-office services. However, some initial applications yielded disappointing results, as managers were unaware of the level of development of the robots' artificial intelligence systems. This study proposes to adapt the Almere model to estimate the technological acceptance of service robots, which express their gender and personality, whilst assisting consumers. A 2 × 2 (two genders vs. two personalities) between-subjects experiment was conducted with 219 participants. Model estimation with Structural Equation Modelling confirmed seven out of eight hypotheses, and all four scenarios were estimated with Ordinary Least Squares, showing that robot gender and personality affected their technological acceptance.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Dimensionality reduction for dynamic movement primitives and application to bimanual manipulation of clothes

  • Colomé, Adrià
  • Torras, Carme
Dynamic movement primitives (DMPs) are widely used as movement parametrization for learning robot trajectories, because of their linearity in the parameters, rescaling robustness, and continuity. However, when learning a movement with DMPs, a very large number of Gaussian approximations needs to be performed. Adding them up for all joints yields too many parameters to be explored when using reinforcement learning (RL), thus requiring a prohibitive number of experiments/simulations to converge to a solution with a (locally or globally) optimal reward. In this paper, we address the process of simultaneously learning a DMP-characterized robot motion and its underlying joint couplings through linear dimensionality reduction (DR), which will provide valuable qualitative information leading to a reduced and intuitive algebraic description of such motion. The results in the experimental section not only show that we can effectively perform DR on DMPs while learning, but we can also obtain better learning curves, as well as additional information about each motion: linear mappings relating joint values and some latent variables., This work was partially developed in the context of the Advanced Grant CLOTHILDE (”CLOTH manIpulation Learning from DEmonstrations”), which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 741930). This work is also partially funded by CSIC projects MANIPlus (201350E102) and TextilRob (201550E028), and Chist-Era Project I-DRESS (PCIN-2015-147)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Exploiting symmetries in reinforcement learning of bimanual robotic tasks

  • Amadio, Fabio
  • Colomé, Adrià
  • Torras, Carme
Movement primitives (MPs) have been widely adopted for representing and learning robotic movements using reinforcement learning policy search. Probabilistic movement primitives (ProMPs) are a kind of MP based on a stochastic representation over sets of trajectories, able to capture the variability allowed while executing a movement. This approach has proved effective in learning a wide range of robotic movements, but it comes with the necessity of dealing with a high-dimensional space of parameters. This may be a critical problem when learning tasks with two robotic manipulators, and this work proposes an approach to reduce the dimension of the parameter space based on the exploitation of symmetry. A symmetrization method for ProMPs is presented and used to represent two movements, employing a single ProMP for the first arm and a symmetry surface that maps that ProMP to the second arm. This symmetric representation is then adopted in reinforcement learning of bimanual tasks (from user-provided demonstrations), using relative entropy policy search algorithm. The symmetry-based approach developed has been tested in an experiment of cloth manipulation, showing a speed increment in learning the task., This work was partially developed in the context of the project CLOTHILDE (>CLOTH manIpulation Learning from DEmonstrations>), which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Advanced Grant agreement No 741930). This work is supported by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Assistive robotics: Research challenges and ethics education initiatives, Robótica asistencial: retos de investigación e iniciativas educativas en roboética

  • Torras, Carme
[EN]Assistive robotics is a fast growing field aimed at helping healthcarers in hospitals, rehabilitation centers and nursery homes, as well as empowering people with reduced mobility at home, so that they can autonomously fulfill their daily living activities. The need to function in dynamic human-centered environments poses new research challenges: robotic assistants need to have friendly interfaces, be highly adaptable and customizable, very compliant and intrinsically safe to people, as well as able to handle deformable materials. Besides technical challenges, assistive robotics raises also ethical defies, which have led to the emergence of a new discipline: Roboethics. Several institutions are developing regulations and standards, and many ethics education initiatives include contents on human-robot interaction and human dignity in assistive situations. In this paper, the state of the art in assistive robotics is briefly reviewed, and educational materials from a university course on Ethics in Social Robotics and AI focusing on the assistive context are presented., [ES] La robótica asistencial es un campo en rápido crecimiento des-tinado a ayudar a los cuidadores en hospitales, centros de rehabilitación y residencias, así como a capacitar a las personas con movilidad reducida en el hogar, para que puedan realizar de forma autónoma sus activida-des cotidianas. La necesidad de desempeñarse en entornos dinámicos centrados en el ser humano plantea nuevos retos de investigación: los asistentes robóticos deben tener interfaces amigables, ser altamente adaptables y personalizables, muy compatibles y seguros para las perso-nas, así como ser capaces de manejar materiales deformables.Además de los desafíos técnicos, la robótica asistencial plantea también desafíos éticos, que han llevado a la aparición de una nueva disciplina: la roboética. Numerosas instituciones están elaborando reglamentos y normas, y muchas iniciativas de educación en ética incluyen contenidos sobre la interacción humano-robot y la dignidad humana en situaciones de asistencia.En este trabajo se revisa brevemente el estado del arte de la robótica asistencial y se presentan materiales educativos de un curso universitario sobre Ética de la Robótica Social e IA en el contexto asistencial., This work is partly supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme through the project CLOTHIL-DE - CLOTH manIpulation Learning from DEmonstrations (Advanced Grant agreement No 741930), and the Spanish Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656)
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Science-Fiction: A Mirror for the Fututre of Humankind, The engaging debate on AI and ethics favored by science-fiction

  • Torras, Carme
Digital technologies have become part of our everyday lives and are increasingly acting as intermediaries in our workplaces and personal relationships or even substituting them. The Internet of things, social networks, programs that learn by interacting with humans, assistive and companion robots, computer games with a purpose, serious games for social impact, roboadvisors, webs that offer digital immortality… These tools can, in a short time, modify the job market, flip someone’s reputation, transform a district, change our relationships —not just at work, but also within our families and close contacts— or extend what a person leaves behind after dying, which now includes a digital footprint., This work has been partly supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme through the project CLOTHILDE – CLOTH manIpulation Learning from DEmonstrations (Advanced Grant agreement No 741930), and the Spanish Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

A Grasping-Centered Analysis for Cloth Manipulation

  • Borràs, Julia
  • Alenyà, Guillem
  • Torras, Carme
Compliant and soft hands have gained a lot of attention in the past decade because of their ability to adapt to the shape of the objects, increasing their effectiveness for grasping. However, when it comes to grasping highly flexible objects such as textiles, we face the dual problem: it is the object that will adapt to the shape of the hand or gripper. In this context, the classic grasp analysis or grasping taxonomies are not suitable for describing textile objects grasps. This work proposes a novel definition of textile object grasps that abstracts from the robotic embodiment or hand shape and recovers concepts from the early neuroscience literature on hand prehension skills. This framework enables us to identify what grasps have been used in literature until now to perform robotic cloth manipulation, and allows for a precise definition of all the tasks that have been tackled in terms of manipulation primitives based on regrasps. In addition, we also review what grippers have been used. Our analysis shows how the vast majority of cloth manipulations have relied only on one type of grasp, and at the same time we identify several tasks that need more variety of grasp types to be executed successfully. Our framework is generic, provides a classification of cloth manipulation primitives and can inspire gripper design and benchmark construction for cloth manipulation., The research leading to these results receives funding from the European Research Council (ERC) from the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE: CLOTH manIpulation Learning from DEmonstrations) and is also supported by the Spanish State Research Agency through the Mar´ıa de Maeztu Seal of Excellence to IRI (MDM-2016-0656) and the “Ramon y Cajal” Fellowship RYC-2017-22703., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Perception of cloth in assistive robotic manipulation tasks

  • Torras, Carme
  • Jiménez Schlegl, Pablo
Assistive robots need to be able to perform a large number of tasks that imply some type of cloth manipulation. These tasks include domestic chores such as laundry handling or bed-making, among others, as well as dressing assistance to disabled users. Due to the deformable nature of fabrics, this manipulation requires a strong perceptual feedback. Common perceptual skills that enable robots to complete their cloth manipulation tasks are reviewed here, mainly relying on vision, but also resorting to touch and force. The use of such basic skills is then examined in the context of the different cloth manipulation tasks, be them garment-only applications in the line of performing domestic chores, or involving physical contact with a human as in dressing assistance., This work was developed in the context of the project CLOTHILDE (“CLOTH manIpulation Learning from DEmonstrations”), which has received funding from the European Research Council (ERC) under the European Union’s Horizon2020 research and innovation programme (Advanced Grant Agreement No. 741930). This work is supported by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

A topological extension of movement primitives for curvature modulation and sampling of robot motion

  • Colomé, Adrià
  • Torras, Carme
This paper proposes to enrich robot motion data with trajectory curvature information. To do so, we use an approximate implementation of a topological feature named writhe, which measures the curling of a closed curve around itself, and its analog feature for two closed curves, namely the linking number. Despite these features have been established for closed curves, their definition allows for a discrete calculation that is well-defined for non-closed curves and can thus provide information about how much a robot trajectory is curling around a line in space. Such lines can be predefined by a user, observed by vision or, in our case, inferred as virtual lines in space around which the robot motion is curling. We use these topological features to augment the data of a trajectory encapsulated as a Movement Primitive (MP). We propose a method to determine how many virtual segments best characterize a trajectory and then find such segments. This results in a generative model that permits modulating curvature to generate new samples, while still staying within the dataset distribution and being able to adapt to contextual variables., Thiswork has been carried out within the projectCLOTHILDE (“CLOTH manIpulation Learning from DEmonstrations”) funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Advanced Grant agreementNo 741930).Research at IRI is also supported by the Spanish State Research Agency through theMaría de Maeztu Seal of Excellence to IRI MDM-2016-0656.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Mobile Manipulation Hackathon: Moving into Real World Applications

  • Roa, Máximo A.
  • Dogar, Mehmet
  • Pagès, Jordi
  • Vivas, Carlos
  • Morales, Antonio
  • Correll, Nikolaus
  • Gorner, Michael
  • Rosell, Jan
  • Foix, Sergi
  • Memmesheimer, Raphael
  • Ferro, Francesco
The Mobile Manipulation Hackathon was held in late 2018 during the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) to showcase the latest applications of wheeled robotic manipulators. The challenge had an open format, where teams developed an application using simulation tools and integrated it into a robotic platform. This article presents the competition and analyzes the results, with information gathered during the event and from a survey circulated among the finalist teams. We provide an overview of the mobile manipulation field, identify key areas required for further development to facilitate the implementation of mobile manipulators in real applications, and discuss ideas about how to structure future hackathon-style competitions to enhance their impact on the scientific and industrial communities., M. Dogar received funding from the UK Engineering and Physical Sciences Research Council under grant EP/P019560/1. A. Morales and the UJI Robotic Intelligence Laboratory are partially funded by Ministerio de Econom´ıa y Competitividad (DPI2017-89910-R) and by Generalitat Valenciana (PROMETEO/2020/034). The work developed by team IOC-AUDECO was partially supported by the Spanish Government through the project DPI2016-80077-R. Team IRI was supported by the European Research Council (ERC) from the European Union H2020 Programme under grant 741930 (CLOTHILDE: CLOTH manIpulation Learning from DEmonstrations), and the Spanish State Research Agency through the Mar´ıa de Maeztu Seal of Excellence to IRI (MDM-2016-0656). Team TAMS was partially funded by the German Research Foundation (DFG) and the National Science Foundation of China in project Crossmodal Learning, TRR-169.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Task-adaptive robot learning from demonstration with gaussian process models under replication

  • Arduengo, Miguel
  • Colomé, Adrià
  • Borràs, Julia
  • Sentis, Luis
  • Torras, Carme
Learning from Demonstration (LfD) is a paradigm that allows robots to learn complex manipulation tasks that can not be easily scripted, but can be demonstrated by a human teacher. One of the challenges of LfD is to enable robots to acquire skills that can be adapted to different scenarios. In this letter, we propose to achieve this by exploiting the variations in the demonstrations to retrieve an adaptive and robust policy, using Gaussian Process (GP) models. Adaptability is enhanced by incorporating task parameters into the model, which encode different specifications within the same task. With our formulation, these parameters can be either real, integer, or categorical. Furthermore, we propose a GP design that exploits the structure of replications, i.e., repeated demonstrations with identical conditions within data. Our method significantly reduces the computational cost of model fitting in complex tasks, where replications are essential to obtain a robust model. We illustrate our approach through several experiments on a handwritten letter demonstration dataset., This work has been partially funded by the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE) and by the Spanish State Research Agency through the Maria de Maeztu Seal of Excellence to IRI [MDM-2016-0656].
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Topological representation of cloth state for robot manipulation: Deriving the configuration space of a rectangular cloth

  • Strazzeri, Fabio
  • Torras, Carme
Forty years ago the notion of configuration space (C-space) revolutionised robot motion planning for rigid and articulated objects. Despite great progress, handling deformable materials has remained elusive because of their infinite-dimensional shape-state space. Finding low-complexity representations has become a pressing research goal. This work tries to make a tiny step in this direction by proposing a state representation for textiles relying on the C-space of some distinctive points. A stratification of the configuration space for n points in the cloth is derived from that of the flag manifold, and topological techniques to determine adjacencies in manipulation-centred state graphs are developed. Their algorithmic implementation permits obtaining cloth state–space representations of different granularities and tailored to particular purposes. An example of their usage to distinguish between cloth states having different manipulation affordances is provided. Suggestions on how the proposed state graphs can serve as a common ground to link the perception, planning and manipulation of textiles are also made., This work is supported by the European Research Council (ERC) within the European Union Horizon 2020 Programme under grant agreement ERC-2016-ADG-741930 (CLOTHILDE: CLOTH manIpulation Learning from DEmonstrations) and by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656). We thank Dr Jordi Sanchez Riera for providing us with cloth simulations.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Do I Have a Personality? Endowing Care Robots with Context-Dependent Personality Traits

  • Andriella, Antonio
  • Siqueira, Henrique
  • Fu, Di
  • Magg, Sven
  • Barros, Pablo
  • Wermter, Stefen
  • Torras, Carme
  • Alenyà, Guillem
Recent studies have revealed the key importance of modelling personality in robots to improve interaction quality by empowering them with social-intelligence capabilities. Most research relies on verbal and non-verbal features related to personality traits that are highly context-dependent. Hence, analysing how humans behave in a given context is crucial to evaluate which of those social cues are effective. For this purpose, we designed an assistive memory game, in which participants were asked to play the game obtaining support from an introvert or extroverted helper, whether from a human or robot. In this context, we aim to (i) explore whether selective verbal and non-verbal social cues related to personality can be modelled in a robot, (ii) evaluate the efficiency of a statistical decision-making algorithm employed by the robot to provide adaptive assistance, and (iii) assess the validity of the similarity attraction principle. Specifically, we conducted two user studies. In the human–human study (N=31), we explored the effects of helper’s personality on participants’ performance and extracted distinctive verbal and non-verbal social cues from the human helper. In the human–robot study (N=24), we modelled the extracted social cues in the robot and evaluated its effectiveness on participants’ performance. Our findings showed that participants were able to distinguish between robots’ personalities, and not between the level of autonomy of the robot (Wizard-of-Oz vs fully autonomous). Finally, we found that participants achieved better performance with a robot helper that had a similar personality to them, or a human helper that had a different personality., This work is supported by the German Research Foundation (DFG), National Natural Science Foundation of China (NSFC), under project Transregio Crossmodal Learning (DFG TRR 169/NSFC 61621136008), and a CAS-DAAD joint fellowship. This project has also been partially funded by the European Union’s Horizon 2020 research and innovation programme under the Marie Skł, odowska-Curie grant agreement SOCRATES (no. 721619), by the European Union’s Horizon 2020 under grant agreement CLOTHILDE (no. 741930), by the Spanish Ministry of Science and Innovation HuMoUR (TIN2017-90086-R), and by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

An inextensible model for the robotic manipulation of textiles

  • Coltraro, Franco
  • Amorós, Jaume
  • Alberich-Carramiñana, Maria
  • Torras, Carme
We introduce a new isometric strain model for the study of the dynamics of cloth garments in a moderate stress environment, such as robotic manipulation in the neighborhood of humans. This model treats textiles as surfaces that are inextensible, admitting only isometric motions. Inextensibility is derived in a continuous setting, prior to any discretization, which gives consistency with respect to remeshing and prevents the problem of locking even with coarse meshes. The simulations of robotic manipulation using the model are compared to the actual manipulation in the real world, finding that the difference between the simulated and the real position of each point in the garment is lower than 1cm in average even when a coarse mesh is used. Aerodynamic contributions to motion are incorporated to the model through the virtual uncoupling of the inertial and gravitational mass of the garment. This approach results in an accurate, when compared to the recorded dynamics of real textiles, description of cloth motion incorporating aerodynamic effects by using only two parameters., This project was developed in the context of the project CLOTHILDE (”CLOTH manIpulation Learning from DEmonstrations”) which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement No. 741930) and is also supported by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656). Maria Alberich-Carramiñana is also with the Barcelona Graduate School of Mathematics (BGSMath) and the Institut de Matemàtiques de la UPC-BarcelonaTech (IMTech). Maria Alberich-Carramiñana and Jaume Amorós are partially supported by the Spanish Research Agency PID2019-103849GB-I00 / AEI / 10.13039/501100011033 grant, and by Generalitat de Catalunya 2017SGR-932 grant.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Household Cloth Object Set: Fostering Benchmarking in Deformable Object Manipulation

  • García-Camacho, Irene
  • Borràs, Julia
  • Calli, Berk
  • Norton, Adam
  • Alenyà, Guillem
Benchmarking of robotic manipulations is one of the open issues in robotic research. An important factor that has enabled progress in this area in the last decade is the existence of common object sets that have been shared among different research groups. However, the existing object sets are very limited when it comes to cloth-like objects that have unique particularities and challenges. This letter is a first step towards the design of a cloth object set to be distributed among research groups from the robotics cloth manipulation community. We present a set of household cloth objects and related tasks that serve to expose the challenges related to gathering such an object set and propose a roadmap to the design of common benchmarks in cloth manipulation tasks, with the intention to set the grounds for a future debate in the community that will be necessary to foster benchmarking for the manipulation of cloth-like objects. Some RGB-D and object scans are collected as examples for the relevant configurations and shared in http://www.iri.upc.edu/groups/perception/ClothObjectSet/., The work of IGC is partially supported by MCIN/ AEI /10.13039/501100011033 and by the ”European Union” under project BURG (PCI2019-103447); JB thanks MCIN/ AEI /10.13039/501100011033 under project CHLOE-GRAPH (PID2020-118649RB-I00); GA acknowledges the EU H2020 Programme under grant agreement ERC–2016–ADG–741930 (CLOTHILDE); and BC thanks the National Science Foundation under grant IIS-1900953
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

How do consumers' gender and rational thinking affect the acceptance of entertainment social robots?

  • Forgas Coll, Santiago
  • Alenyà, Guillem
  • Huertas García, Rubén
  • Andriella, Antonio
In recent years, the rapid ageing of the population, a longer life expectancy and elderly people¿s desire to live independently are social changes that put pressure on healthcare systems. This context is boosting the demand for companion and entertainment social robots on the market and, consequently, producers and distributors are interested in knowing how these social robots are accepted by consumers. Based on technology acceptance models, a parsimonious model is proposed to estimate the intention to use this new advanced social robot technology and, in addition, an analysis is performed to determine how consumers¿ gender and rational thinking condition the precedents of the intention to use. The results show that gender differences are more important than suggested by the literature. While women gave greater social influence and perceived enjoyment as the main motives for using a social robot, in contrast, men considered their perceived usefulness to be the principal reason and, as a differential argument, the ease of use. Regarding the reasoning system, the most significant differences occurred between heuristic individuals, who stated social influence as the main reason for using a robot, and the more rational consumers, who gave ease of use as a differential argument., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work has been partially funded by the EU Project SOCRATES H2020-MSCA-ITN-721619, by the European Union's Horizon 2020 under ERC Advanced Grant CLOTHILDE (No. 741930), by the Spanish Ministry of Science and Innovation HuMoUR TIN2017-90086-R, and by the State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

“I know how you feel” The importance of interaction style on users’ acceptance in an entertainment scenario

  • Andriella, Antonio
  • Huertas García, Rubén
  • Forgas Coll, Santiago
  • Torras, Carme
  • Alenyà, Guillem
In this article, we aim to evaluate the role of robots’ personality-driven behavioural patterns on users’ intention to use in an entertainment scenario. Toward such a goal, we designed two personalities: one introverted with an empathic and self-comparative interaction style, and the other extroverted with a provocative and other-comparative interaction style. To evaluate the proposed technology acceptance model, we conducted an experiment (N=209) at a public venue where users were requested to play a game with the support of the TIAGo robot. Our findings show that the robot personality affects the acceptance model and three relevant drivers: perceived enjoyment, perceived usefulness, and social influence. The extroverted robot was perceived as more useful than the introverted, and participants who interacted with it were faster at solving the game. On the other hand, the introverted robot was perceived as more enjoyable but less useful than the extroverted, and participants who interacted with it made fewer mistakes. Taken together, these findings support the importance of designing proper robot personalities in influencing users’ acceptance, featuring that a given style can elicit a different driver of acceptance., *This work has been partially funded by the European Union's Horizon 2020 under ERC Advanced Grant CLOTHILDE (no. 741930); by MCIN/ AEI /10.13039/501100011033 and by the "European Union NextGen- erationEU/PRTR under the project ROB-IN (PLEC2021-007859) and the project COHERENT (PCI2020- 120718-2); and by the European Commission{ NextGenerationEU, through CSIC's Thematic Platforms (PTI+ Neuro-Aging).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

The effects of gender and personality of robot assistants on customers’ acceptance of their service

  • Forgas Coll, Santiago
  • Huertas García, Rubén
  • Andriella, Antonio
  • Alenyà, Guillem
The Covid-19 pandemic has stimulated the use of social robots in front-office services. However, some initial applications yielded disappointing results, as managers were unaware of the level of development of the robots’ artificial intelligence systems. This study proposes to adapt the Almere model to estimate the technological acceptance of service robots, which express their gender and personality, whilst assisting consumers. A 2 × 2 (two genders vs. two personalities) between-subjects experiment was conducted with 219 participants. Model estimation with Structural Equation Modelling confirmed seven out of eight hypotheses, and all four scenarios were estimated with Ordinary Least Squares, showing that robot gender and personality affected their technological acceptance., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work has also been partially funded by the EU project SOCRATES H2020-MSCA-ITN-721619, by the European Union’s Horizon 2020 under ERC-Advanced Grant CLOTHILDE (No. 741930), by the Spanish Ministry of Science and Innovation HuMoUR TIN2017-90086-R and by the State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Mixtures of controlled Gaussian processes for dynamical modeling of deformable objects

  • Zheng, Ce Xu
  • Colomé, Adrià
  • Sentis, Luis
  • Torras, Carme
Trabajo presentado en el 4th Annual Conference on Learning for Dynamics and Control, celebrado en El Camino Real, Palo Alto, CA (Estados Unidos), los días 23 y 24 de junio de 2022, Control and manipulation of objects is a highly relevant topic in Robotics research. Although significant advances have been made over the manipulation of rigid bodies, the manipulation of non-rigid objects is still challenging and an open problem. Due to the uncertainty of the outcome when applying physical actions to non-rigid objects, using prior knowledge on objects¿ dynamics can greatly improve the control performance. However, fitting such models is a challenging task for materials such as clothing, where the state is represented by points in a mesh, resulting in very large dimensionality that makes models difficult to learn, process and predict based on measured data. In this paper, we expand previous work on Controlled Gaussian Process Dynamical Models (CGPDM), a method that uses a non-linear projection of the state space onto a much smaller dimensional latent space, and learns the object dynamics in the latent space. We take advantage of the variability in training data by employing Mixture of Experts (MoE), and we devise theory and experimental validations that demonstrate significant improvements in training and prediction times, plus robustness and error stability when predicting deformable objects exposed to disparate movement ranges., This work was partially developed in the context of the project CLOTHILDE (”CLOTH manIpulation Learning from DEmonstrations”), which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Advanced Grant agreement No 741930). We would like to thank the members of the HCRL Lab and the Department of Aerospace Engineering and Engineering Mechanics at The University of Texas at Austin for their feedback during the development of this work.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Evaluating the effect of Theory of Mind on people's trust in a faulty robot

  • Rossi, Alessandra
  • Andriella, Antonio
  • Rossi, Silvi
  • Torras, Carme
  • Alenyà, Guillem
Trabajo presentado en el 31st IEEE International Conference on Robot and Human Interactive Communication (RO-MAN), celebrado en Nápoles (Italia), del 29 de agosto al 2 de septiembre de 2022, The success of human-robot interaction is strongly affected by the people's ability to infer others' intentions and behaviours, and the level of people's trust that others will abide by their same principles and social conventions to achieve a common goal. The ability of understanding and reasoning about other agents' mental states is known as Theory of Mind (ToM). ToM and trust, therefore, are key factors in the positive outcome of human-robot interaction. We believe that a robot endowed with a ToM is able to gain people's trust, even when this may occasionally make errors. In this work, we present a user study in the field in which participants (N=123) interacted with a robot that may or may not have a ToM, and may or may not exhibit erroneous behaviour. Our findings indicate that a robot with ToM is perceived as more reliable, and they trusted it more than a robot without a ToM even when the robot made errors. Finally, ToM results to be a key driver for tuning people's trust in the robot even when the initial condition of the interaction changed (i.e., loss and regain of trust in a longer relationship)., This work has been supported by Italian PON R&I 2014-2020 - REACT-EU (CUP E65F21002920003); by the European Union’s Horizon 2020 under ERC Advanced Grant CLOTHILDE (no. 741930); by MCIN/ AEI /10.13039/501100011033 and by the ”European Union NextGenerationEU/PRTR” under the project ROBIN (PLEC2021-007859) and the project COHERENT (PCI2020-120718-2); by the ”European Union NextGenerationEU/PRTR” through CSIC’s Thematic Platforms (PTI+ Neuro-Aging) and by the European Union’s Horizon 2020 research and innovation programme PRO-CARED (no.801342) under Marie Skłodowska-Curie grant agreement (Tecniospring INDUSTRY) and the Government of Catalonia’s Agency for Business Competitiveness (ACCIO).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Garment manipulation dataset for robot learning by demonstration through a virtual reality framework

  • Boix Granell, Arnau
  • Foix, Sergi
  • Torras, Carme
Being able to teach complex capabilities, such as folding garments, to a bi-manual robot is a very challenging task, which is often tackled using learning from demonstration datasets. The few garment folding datasets available nowadays to the robotics research community are either gathered from human demonstrations or generated through simulation. The former have the huge problem of perceiving human action and transferring it to the dynamic control of the robot, while the latter requires coding human motion into the simulator in open loop, resulting in far-from-realistic movements. In this article, we present a reduced but very accurate dataset of human cloth folding demonstrations. The dataset is collected through a novel virtual reality (VR) framework we propose, based on Unity¿s 3D platform and the use of a HTC Vive Pro system. The framework is capable of simulating very realistic garments while allowing users to interact with them, in real time, through handheld controllers. By doing so, and thanks to the immersive experience, our framework gets rid of the gap between the human and robot perception-action loop, while simplifying data capture and resulting in more realistic samples., This work was developed in the context of the project CLOTHILDE (”CLOTH manIpulation Learning from DEmonstrations”) which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement No. 741930) and is also supported by the BURG project PCI2019-103447 funded by MCIN/ AEI /10.13039/501100011033 and by the ”European Union”.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Implications of robot backchannelling in cognitive therapy

  • Andriella, Antonio
  • Torras, Carme
  • Alenyà, Guillem
Trabajo presentado en la 14th International Conference on Social Robotics , celebrada en Florencia (Italia), del 13 al 16 de diciembre de 2022, The social ability of humans to provide active feedback during conversations is known as backchannelling. Recent work has recognised the importance of endowing robots with such social behaviour to make interactions more natural. Nonetheless, very little is known about how backchannelling should be designed in order to be detected and whether it can have an impact on users' behaviour and performance in cooperative tasks. In this article, we aim at evaluating the legibility of robot's backchannelling behaviour on Persons with Dementia (PwDs) and its effect on their performance when playing cognitive training exercises. Aiming to do so, a TIAGo robot was endowed with backchannelling behaviour generated by combining verbal and non-verbal cues. To evaluate our system, two user studies were carried out, in which the social signal was provided first by a human therapist and later on by a robot. Results indicate that patients were capable of identifying such kind of feedback. Nonetheless, our findings pointed out a significant difference in terms of performance between the two studies. They reveal how patients in the study with the robot overused the feedback to obtain the correct answer, putting in place a cheating mechanism that has led them to significantly worsen their performance. We conclude our work by discussing the implications of our findings when deploying robots in sensitive roles and possible solutions to address such unexpected behaviours., This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 801342 (Tecniospring INDUSTRY). This project has also been supported by the European Union’s Horizon 2020 under ERC Advanced Grant CLOTHILDE (no. 741930); by MCIN/AEI/10.13039/501100011033 by the “European Union NextGenerationEU/PRTR” under the project ROB-IN (PLEC2021-007859); by the Research Council of Norway under the project SECUROPS (INT-NO/0875); and by the “European Union NextGenerationEU/PRTR” through CSIC’s Thematic Platforms (PTI+ Neuro-Aging).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Ordinal inverse reinforcement learning applied to robot learning with small data

  • Colomé, Adrià
  • Torras, Carme
Trabajo presentado en el International Conference on Intelligent Robots and Systems (IROS), celebrado en Kyoto (Japón), del 23 al 27 de octubre de 2022, Over the last decade, the ability to teach actions to robots in a user-friendly way has gained relevance, and a practical way of teaching robots a new task is to use Inverse Reinforcement Learning (IRL). In IRL, an expert teacher shows the robot a desired behaviour and an agent builds a model of the reward. The agent can also infer a policy that performs in an optimal way within the limitations of the knowledge provided to it. However, most IRL approaches assume an (almost) optimal performance of the teaching agent, which might become unpractical if the teacher is not actually an expert. In addition, most IRL focus on discrete state-action spaces that limit their applicability to certain real-world problems such as within the context of direct Policy Search (PS) reinforcement learning. Therefore, in this paper we introduce Ordinal Inverse Reinforcement Learning (OrdIRL) for continuous state variables, in which the teacher can qualitatively evaluate robot performance by selecting one among the predefined performance levels (e.g. bad, medium, goodu for three tiers of performance). Once the OrdIRL has fit an ordinal distribution to the data, we propose to use Bayesian Optimization (BO) to either gain knowledge on the inferred model (exploration) or find a policy or action that maximizes the expected reward given the prior knowledge on the reward (exploitation). In the case of large-dimensional state-action spaces, we use Dimensionality Reduction (DR) techniques and perform the BO in the latent space. Experimental results on simulation and with a robot arm show how this approach allows for learning the reward function with small data., CLOTHILDE - CLOTH manIpulation Learning from DEmonstrations (EC-H2020-741930), Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Perturbation-Based stiffness inference in variable impedance control

  • Caldarelli, Edoardo
  • Colomé, Adrià
  • Torras, Carme
One of the major challenges in learning from demonstration is to teach the robot a wider set of task features than the plain trajectories to be followed. In this sense, one key parameter is stiffness, i.e., the rigidity that the manipulator should exhibit when performing a task. The required robot stiffness is often not known a priori and varies along the execution of the task, thus its profile needs to be inferred from the demonstrations. In this work, we propose a novel, force-based algorithm for inferring time-varying stiffness profiles, leveraging the relationship between stiffness and tracking error, and involving human-robot interaction. We begin by gathering a set of demonstrations with kinesthetic teaching. Then, the robot executes a perturbed reference, obtained from these demonstrations by means of Gaussian process regression, and the human intervenes if the perturbation makes the manipulator deviate from its expected behaviour. Human intervention is measured and used to infer the desired control stiffness. In the experiments section, we show that our algorithm can be combined with different types of force sensors, and provide suitable processing algorithms. Our approach correctly infers the stiffness profiles from the force and electromyography sensors, their combination permitting also to comply with the physical constraints imposed by the environment. This is demonstrated in three experiments of increasing complexity: a motion in free Cartesian space, a rigid assembly task, and bed-making., This paper was recommended for publication by Editor Jens Kober upon evaluation of the Associate Editor and Reviewers’ comments. This work was supported by the project CLOTHILDE (“CLOTH manIpulation Learning from DEmonstrations”), funded from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Advanced Grant agreement No 741930).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Does the Personality of Consumers Influence the Assessment of the Experience of Interaction with Social Robots?

  • Forgas Coll, Santiago
  • Huertas García, Rubén
  • Andriella, Antonio
  • Alenyà, Guillem
In recent years, in response to the effects of Covid-19, there has been an increase in the use of social robots in service organisations, as well as in the number of interactions between consumers and robots. However, it is not clear how consumers are valuing these experiences or what the main drivers that shape them are. Furthermore, it is an open research question whether these experiences undergone by consumers can be affected by their own personality. This study attempts to shed some light on these questions and, to do so, an experiment is proposed in which a sample of 378 participants evaluate a simulated front-office service experience delivered by a social robot. The authors investigate the underlying process that explains the experience and find that cognitive-functional factors, emphasising efficiency, have practically the same relevance as emotional factors, emphasising stimulation. In addition, this research identifies the personality traits of the participants and explores their moderating role in the evaluation of the experience. The results reveal that each personality trait, estimated between high and low poles, generates different responses in the evaluation of the experience., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work has been partially funded by the European Union's Horizon 2020 under ERC Advanced Grant CLOTHILDE (no. 741930); by MCIN/ AEI/10.13039/501100011033 and by the "European Union NextGenerationEU/PRTR" under the project ROB-IN (PLEC2021-007859) and the project COHERENT (PCI2020-120718-2); and by the "European Union NextGenerationEU/PRTR" through CSIC's Thematic Platforms (PTI + Neuro-Aging)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

A Representation of Cloth States based on a Derivative of the Gauss Linking Integral

  • Coltraro, Franco
  • Fontana, Josep
  • Amorós, Jaume
  • Alberich-Carramiñana, Maria
  • Borràs, Julia
  • Torras, Carme
Robotic manipulation of cloth is a complex task because of the infinite-dimensional shape-state space of textiles, which makes their state estimation very difficult. In this paper we introduce the dGLI Cloth Coordinates, a finite low-dimensional representation of cloth states that allows us to efficiently distinguish a large variety of different folded states, opening the door to efficient learning methods for cloth manipulation planning and control. Our representation is based on a directional derivative of the Gauss Linking Integral and allows us to represent spatial as well as planar folded configurations in a consistent and unified way. The proposed dGLI Cloth Coordinates are shown to be more accurate in the representation of cloth states and significantly more sensitive to changes in grasping affordances than other classic shape distance methods. Finally, we apply our representation to real images of a cloth, showing that with it we can identify the different states using a distance-based classifier., This work was developed under the project CLOTHILDE which has received funding from the European Research Council (ERC) under the EU-Horizon 2020 research and innovation programme (grant agreement No. 741930). M. Alberich-Carramiñana is also with the Barcelona Graduate School of Mathematics (BGSMath) and the Institut de Matemàtiques de la UPC-BarcelonaTech (IMTech), and she and J. Amorós are partially supported by the Spanish State Research Agency AEI/10.13039/501100011033 grant PID2019-103849GB-I00 and by the AGAUR project 2021 SGR 00603 Geometry of Manifolds and Applications, GEOMVAP. J. Borràs is supported by the Spanish State Research Agency MCIN/ AEI /10.13039/501100011033 grant PID2020-118649RB-I00 (CHLOE-GRAPH project).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Generating predicate suggestions based on the space of plans: an example of planning with preferences

  • Canal, Gerard
  • Torras, Carme
  • Alenyà, Guillem
Task planning in human–robot environments tends to be particularly complex as it involves additional uncertainty introduced by the human user. Several plans, entailing few or various differences, can be obtained to solve the same given task. To choose among them, the usual least-cost plan criteria is not necessarily the best option, because here, human constraints and preferences come into play. Knowing these user preferences is very valuable to select an appropriate plan, but the preference values are usually hard to obtain. In this context, we propose the Space-of-Plans-based Suggestions (SoPS) algorithms that can provide suggestions for some planning predicates, which are used to define the state of the environment in a task planning problem where actions modify the predicates. We denote these predicates as suggestible predicates, of which user preferences are a particular case. The first algorithm is able to analyze the potential effect of the unknown predicates and provide suggestions to values for these unknown predicates that may produce better plans. The second algorithm is able to suggest changes to already known values that potentially improve the obtained reward. The proposed approach utilizes a Space of Plans Tree structure to represent a subset of the space of plans. The tree is traversed to find the predicates and the values that would most increase the reward, and output them as a suggestion to the user. Our evaluation in three preference-based assistive robotics domains shows how the proposed algorithms can improve task performance by suggesting the most effective predicate values first., This work has been partially supported by the ERC project Clothilde (ERC-2016-ADG-741930); by MCIN/AEI/10.13039/501100011033 under the project CHLOE-GRAPH (PID2020-118649RB-l00); by the European Union NextGenerationEU/PRTR under the project ROB-IN (PLEC2021-007859); and by the CHIST-ERA project COHERENT (EPSRC EP/V062506/1). Gerard Canal has also been supported by the Spanish Ministry of Education, Culture and Sport by the FPU15/00504 doctoral grant and by the Royal Academy of Engineering and the Office of the Chief Science Adviser for National Security under the UK Intelligence Community Postdoctoral Research Fellowship programme., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

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

  • Andriella, Antonio
  • Torras, Carme
  • Abdelnour, Carla
  • Alenyà, Guillem
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 as domain experts have not been fully involved in the entire pipeline of the design process as well as in the automatisation of the robots’ behaviour. In this article, we present aCtive leARning agEnt aSsiStive bEhaviouR (CARESSER), a novel framework that actively learns robotic assistive behaviour by leveraging the therapist’s expertise (knowledge-driven approach) and their demonstrations (data-driven approach). By exploiting that hybrid approach, the presented method enables in situ fast learning, in a fully autonomous fashion, of personalised patient-specific policies. With the purpose of evaluating our framework, we conducted two user studies in a daily care centre in which older adults affected by mild dementia and mild cognitive impairment (N = 22) were requested to solve cognitive exercises with the support of a therapist and later on of a robot endowed with CARESSER. Results showed that: (i) the robot managed to keep the patients’ performance stable during the sessions even more so than the therapist; (ii) the assistance offered by the robot during the sessions eventually matched the therapist’s preferences. We conclude that CARESSER, with its stakeholder-centric design, can pave the way to new AI approaches that learn by leveraging human–human interactions along with human expertise, which has the benefits of speeding up the learning process, eliminating the need for the design of complex reward functions, and finally avoiding undesired states., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work has been partially funded by the European Union’s Horizon 2020 under ERC Advanced Grant CLOTHILDE (no. 741930); by MCIN/ AEI /10.13039/501100011033 and by the “European Union NextGenerationEU/PRTR” under the project ROB-IN (PLEC2021-007859) and the project COHERENT (PCI2020-120718-2); and by the “European Union NextGenerationEU/PRTR”, through CSIC’s Thematic Platforms (PTI+ Neuro-Aging).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Social robot-delivered customer-facing services: an assessment of the experience, 由社交机器人提供的面向客户的服务:对体验的分析

  • Forgas Coll, Santiago
  • Huertas García, Rubén
  • Andriella, Antonio
  • Alenyà, Guillem
The ability to install social intelligence protocols in robots in order for them to exhibit conversational skills has made them ideal tools for delivering services with a high cognitive and low emotional load. Little is known about how this capability influences the customer experience and the intention to continue receiving these services. Experiences were assessed in a study simulating customer-facing service delivery, and the constructs of the technology readiness index and stated gender were analysed as possible moderators in a quasi-experiment. Hedonic quality was the most relevant factor explaining attitude, and attitude explained intention to use as well as social influence. As for the constructs of technological readiness and gender, optimism and innovativeness seem to be the most likely candidates for moderating the other variables. The most optimistic and the most innovative route would be for the main actors to continue adapting to social robot technology in the future., This work was partially funded by the European Union’s Horizon 2020 under ERC Advanced Grant CLOTHILDE (no. 741930); by MCIN/AEI/10.13039/501100011033 and by the ‘European Union NextGenerationEU/PRTR’ under the project ROB-IN (PLEC2021-007859); by the Research Council of Norway under the project SECUROPS (INT-NO/0875); and by the ‘European Union NextGenerationEU/PRTR’ through CSIC’s Thematic Platforms (PTI + Neuro-Aging); A. Andriella was supported by the European Union’s Horizon 2020 research and innovation programme PRO-CARED (no.801342) under Marie Skłodowska-Curie grant agreement (Tecniospring INDUSTRY) and the Government of Catalonia’s Agency for Business Competitiveness (ACCIÓ)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Controlled Gaussian process dynamical models with application to robotic cloth manipulation

  • Amadio, Fabio
  • Delgado-Guerrero, Juan Antonio
  • Colomé, Adrià
  • Torras, Carme
Over the last years, significant advances have been made in robotic manipulation, but still, the handling of non-rigid objects, such as cloth garments, is an open problem. Physical interaction with non-rigid objects is uncertain and complex to model. Thus, extracting useful information from sample data can considerably improve modeling performance. However, the training of such models is a challenging task due to the high-dimensionality of the state representation. In this paper, we propose Controlled Gaussian Process Dynamical Models (CGPDMs) for learning high-dimensional, nonlinear dynamics by embedding them in a low-dimensional manifold. A CGPDM is constituted by a low-dimensional latent space, with an associated dynamics where external control variables can act and a mapping to the observation space. The parameters of both maps are marginalized out by considering Gaussian Process priors. Hence, a CGPDM projects a high-dimensional state space into a smaller dimension latent space, in which it is feasible to learn the system dynamics from training data. The modeling capacity of CGPDM has been tested in both a simulated and a real scenario, where it proved to be capable of generalizing over a wide range of movements and confidently predicting the cloth motions obtained by previously unseen sequences of control actions., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work was partially developed in the context of the project CLOTHILDE (“CLOTH manIpulation Learning from DEmonstrations”), which has received funding from ERC under the European Union’s Horizon 2020 research and innovation program (Advanced Grant agreement No 741930). This work has also received funding from project CHLOE-GRAPH (PID2020-118649RB-I00) funded by MCIN/ AEI /10.13039/501100011033., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Gendered Human–Robot Interactions in Services

  • Forgas Coll, Santiago
  • Huertas García, Rubén
  • Andriella, Antonio
  • Alenyà, Guillem
The outbreak of Covid-19 precipitated the use of service robots in customer-facing services as a replacement for employees to avoid human-to-human contact. However, this development has not resolved the debate as to whether robots should be characterized with gender attributes or simply be genderless. This study explores whether endowing a robot with gender attributes makes it more acceptable as a service provider among stated men and women. To this end, an experiment was conducted at a public fair in which a gendered robot simulated the provision of a service to customers, which consisted of offering them advice, hints, and messages of encouragement to help complete a eudaemonic puzzle. A parsimonious version of the Almere model was used to estimate acceptance of the technology. The findings reveal that for both stated men and women, the main drivers for accepting the female-coded robot are perceived usefulness and social influence, although women attach greater importance to social influence. For the male-coded robot, perceived usefulness and social influence are the main arguments for women, while for men they are enjoyment, perceived usefulness and, negatively, ease of use. In addition, different indirect effects between stated sexes are also identified. In summary, men and women consider different factors when accepting robots of each gender., This work was partially funded by the European Union’s Horizon 2020 programme under ERC Advanced Grant CLOTHILDE (no. 741930); and by the European Union’s Horizon 2020 research and innovation programme under the Marie Sk?odowska-Curie grant agreement No 801342 (Tecniospring INDUSTRY); by MCIN/ AEI /https://doi.org/10.13039/501100011033 and by the “European Union NextGenerationEU/PRTR” under the project ROB-IN (PLEC2021-007859); by the Research Council of Norway under the project SECUROPS (INT-NO/0875); and by the “European Union NextGenerationEU/PRTR” through CSIC’s Thematic Platforms (PTI + Neuro-Aging). Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Correction to: Gendered Human–Robot Interactions in Services (International Journal of Social Robotics, (2023), 15, 11, (1791-1807), 10.1007/s12369-023-01035-8)

  • Forgas Coll, Santiago
  • Huertas García, Rubén
  • Andriella, Antonio
  • Alenyà, Guillem
This paper was inadvertently published with incorrect funding information. Please find the incorrect and correct versions below. Springer wishes to apologize for any inconvenience caused. Incorrect funding information: This work was partially funded by the European Union’s Horizon 2020 programme under ERC Advanced Grant CLOTHILDE (no. 741,930) and under the Marie Skłodowska-Curie grant agreement (no. 712,949) (TECNIOspring PLUS); by MCIN/ AEI /https://doi.org/10.13039/501100011033 and by the “European Union NextGenerationEU/PRTR” under the project ROB-IN (PLEC2021-007859); by the Research Council of Norway under the project SECUROPS (INT-NO/0875); and by the “European Union NextGenerationEU/PRTR” through CSIC’s Thematic Platforms (PTI + Neuro-Aging). Correct funding information: This work was partially funded by the European Union’s Horizon 2020 programme under ERC Advanced Grant CLOTHILDE (no. 741,930); and by the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 801,342 (Tecniospring INDUSTRY); by MCIN/ AEI /https://doi.org/10.13039/501100011033 and by the “European Union NextGenerationEU/PRTR” under the project ROB-IN (PLEC2021-007859); by the Research Council of Norway under the project SECUROPS (INT-NO/0875); and by the “European Union NextGenerationEU/PRTR” through CSIC’s Thematic Platforms (PTI + Neuro-Aging)., This work was partially funded by the European Union’s Horizon 2020 programme under ERC Advanced Grant CLOTHILDE (no. 741,930); and by the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 801,342 (Tecniospring INDUSTRY); by MCIN/ AEI /https://doi.org/10.13039/501100011033 and by the “European Union NextGenerationEU/PRTR” under the project ROB-IN (PLEC2021-007859); by the Research Council of Norway under the project SECUROPS (INT-NO/0875); and by the “European Union NextGenerationEU/PRTR” through CSIC’s Thematic Platforms (PTI + Neuro-Aging)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Quadratic Dynamic Matrix Control for Fast Cloth Manipulation

  • Caldarelli, Edoardo
  • Colomé, Adrià
  • Ocampo-Martínez, Carlos
  • Torras, Carme
Trabajo presentado en la International Conference on Intelligent Robots and Systems (IROS), celebrada en Detroit, MI (Estados Unidos), del 1 al 5 de octubre de 2023, Robotic cloth manipulation is an increasingly relevant area of research, challenging classic control algorithms due to the deformable nature of cloth. While it is possible to apply linear model predictive control to make the robot move the cloth according to a given reference, this approach suffers from a large dimensionality of the state-space representation of the cloth models. To address this issue, in this work we study the application of an input-output model predictive control strategy, based on quadratic dynamic matrix control, to robotic cloth manipulation. To account for uncertain disturbances on the cloth's motion, we further extend the algorithm with suitable chance constraints. In extensive simulated experiments, involving disturbances and obstacle avoidance, we show that quadratic dynamic matrix control can be successfully applied in different cloth manipulation scenarios, with significant gains in optimization speed compared to standard model predictive control strategies. The experiments further demonstrate that the closed-loop model used by quadratic dynamic matrix control can be beneficial to the tracking accuracy, leading to improvements over the standard predictive control strategy. Moreover, a preliminary experiment on a real robot shows that quadratic dynamic matrix control can indeed be employed in real settings., This work was supported by the project CLOTHILDE (“CLOTH manIpulation Learning from DEmonstrations”), funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Advanced Grant agreement No 741930)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

A Virtual Reality Framework For Fast Dataset Creation Applied to Cloth Manipulation with Automatic Semantic Labelling

  • Borràs, Julia
  • Boix-Granell, Arnau
  • Foix, Sergi
  • Torras, Carme
Trabajo presentado en la IEEE International Conference on Robotics and Automation (ICRA), celebrada en Londres (UK), del 29 de mayo al 2 de junio de 2024, Teaching complex manipulation skills, such as folding garments, to a bi-manual robot is a very challenging task, which is often tackled through learning from demon-stration. The few datasets of garment-folding demonstrations available nowadays to the robotics research community have been either gathered from human demonstrations or generated through simulation. The former have the great difficulty of perceiving both cloth state and human action as well as transferring them to the dynamic control of the robot, while the latter require coding human motion into the simulator in open loop, i.e., without incorporating the visual feedback naturally used by people, resulting in far-from-realistic movements. In this article, we present an accurate dataset of human cloth folding demonstrations. The dataset is collected through our novel virtual reality (VR) framework, based on Unity's 3D platform and the use of an HTC Vive Pro system. The framework is capable of simulating realistic garments while allowing users to interact with them in real time through handheld controllers. By doing so, and thanks to the immersive experience, our framework permits exploiting human visual feedback in the demonstrations while at the same time getting rid of the difficulties of capturing the state of cloth, thus simplifying data acquisition and resulting in more realistic demonstrations. We create and make public a dataset of cloth manipulation sequences, whose cloth states are semantically labeled in an automatic way by using a novel low-dimensional cloth representation that yields a very good separation between different cloth configurations., The research leading to these results receives funding from the European Research Council (ERC) from the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE: CLOTH manIpulation Learning from DEmonstrations) and project SoftEnable (HORIZONCL4- 2021-DIGITAL-EMERGING-01-101070600). Authors also received funding from project CHLOE-GRAPH (PID2020-118649RB-I00) funded by MCIN/ AEI /10.13039/501100011033 and COHERENT (PCI2020-120718- 2) funded by MCIN/ AEI /10.13039/501100011033 and cofunded by the ”European Union NextGenerationEU/PRTR”., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

A novel collision model for inextensible textiles and its experimental validation

  • Coltraro, Franco
  • Amorós, Jaume
  • Alberich-Carramiñana, Maria
  • Torras, Carme
In this work, we introduce a collision model specifically tailored for the simulation of inextensible textiles. The model considers friction, contacts, and inextensibility constraints all at the same time without any decoupling. Self-collisions are modeled in a natural way that allows considering the thickness of cloth without introducing unwanted oscillations. The discretization of the equations of motion leads naturally to a sequence of quadratic problems with inequality and equality constraints. In order to solve these problems efficiently, we develop a novel active-set algorithm that takes into account past active constraints to accelerate the resolution of unresolved contacts. We put to a test the developed collision procedure with diverse scenarios involving static and dynamic friction, sharp objects, and complex-topology folding sequences. Finally, we perform an experimental validation of the collision model by comparing simulations with recordings of real textiles as given by a Motion Capture System. The results are very accurate, having errors around 1 cm for DIN A2 textiles (42 x 59.4 cm) even in difficult scenarios involving fast and strong hits with a rigid object., This work was developed in the context of the project CLOTHILDE (“CLOTH manIpulation Learning from DEmonstrations”) which has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (grant agreement No. 741930). M. Alberich-Carramiñana is also with the Barcelona Graduate School of Mathematics (BGSMath) and the Institut de Matemàtiques de la UPC-BarcelonaTech (IMTech), and she and J. Amorós are partially supported by the Spanish State Research Agency AEI/10.13039/501100011033 grant PID2019-103849GB-I00 and by the AGAUR project 2021 SGR 00603 Geometry of Manifolds and Applications, GEOMVAP., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Benchmarking the Sim-to-Real Gap in Cloth Manipulation

  • Blanco-Mulero, David
  • Barbany, Oriol
  • Alcan, Gokhan
  • Colomé, Adrià
  • Torras, Carme
  • Kyrki, Ville
Realistic physics engines play a crucial role for learning to manipulate deformable objects such as garments in simulation. By doing so, researchers can circumvent challenges such as sensing the deformation of the object in the real-world. In spite of the extensive use of simulations for this task, few works have evaluated the reality gap between deformable object simulators and real-world data. We present a benchmark dataset to evaluate the sim-to-real gap in cloth manipulation. The dataset is collected by performing a dynamic as well as a quasi-static cloth manipulation task involving contact with a rigid table. We use the dataset to evaluate the reality gap, computational time, and simulation stability of four popular deformable object simulators: MuJoCo, Bullet, Flex, and SOFA. Additionally, we discuss the benefits and drawbacks of each simulator. The benchmark dataset is open-source. Supplementary material, videos, and code, can be found at https://sites.google.com/view/cloth-sim2real-benchmark., This work was supported in part by Academy of Finland under Grant 317020, in part by Business Finland (decision 9249/31/2021), in part by EU’s H2020 Programme through the CLOTHILDE Project under Grant ERC-2016-ADG-741930, in part by EU’s Horizon Europe Programme through the SoftEnable Project under Grant Agreement 101070600, and in part by MCIN/AEI/10.13039/501100011033 through CHLOE-GRAPH Project under Grant PID2020-118649RB-I00., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Deformable surface reconstruction via Riemannian metric preservation

  • Barbany, Oriol
  • Colomé, Adrià
  • Torras, Carme
Estimating the pose of an object from a monocular image is a fundamental inverse problem in computer vision. Due to its ill-posed nature, solving this problem requires incorporating deformation priors. In practice, many materials do not perceptibly shrink or extend when manipulated, constituting a reliable and well-known prior. Mathematically, this translates to the preservation of the Riemannian metric. Neural networks offer the perfect playground to solve the surface reconstruction problem as they can approximate surfaces with arbitrary precision and allow the computation of differential geometry quantities. This paper presents an approach for inferring continuous deformable surfaces from a sequence of images, which is benchmarked against several techniques and achieves state-of-the-art performance without the need for offline training. Being a method that performs per-frame optimization, our method can refine its estimates, contrary to those based on performing a single inference step. Despite enforcing differential geometry constraints at each update, our approach is the fastest of all the tested optimization-based methods., This work is part of the project CLOTHILDE (“CLOTH manIpulation Learning from DEmonstrations”) which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (Advanced Grant agreement No. 741930). O.B. thanks Franco Coltraro and Xavier Gràcia for their insights on differential geometry, and the European Laboratory for Learning and Intelligent Systems (ELLIS) PhD program for support., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Ethics of Social Robotics: Individual and Societal Concerns and Opportunities

  • Torras, Carme
Focus on the ethics of a given technology tends to lag far behind its development. This lag has been particularly acute in the case of artificial intelligence, whose accelerated deployment in a wide range of domains has triggered unprecedented attention on the risks and consequences for society at large, leading to a myriad of ethics regulations, which are difficult to coordinate and integrate due to their late appearance. The very nature of social robots forces their deployment to occur at a much slower pace, providing an opportunity for a profound reflection on ethics, which is already happening in multidisciplinary teams. This article provides a personal view of the ethics landscape, centered on the particularities of social robotics, with the main issues being ordered along two axes (individual and societal) and grouped into eight categories (human dignity, human autonomy, robot transparency, emotional bonding, privacy and safety, justice, freedom, and responsibility). This structure stems from the experience of developing and teaching a university course on ethics in social robotics, whose pedagogical materials are freely available., This work was partly supported by the European Research Council through the project CLOTHILDE (Cloth Manipulation Learning from Demonstrations; advanced grant agreement 741930) and by the European Union’s Horizon Europe program through the project SoftEnable (HORIZON-CL4-2021-DIGITAL-EMERGING-01-101070600)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Model Predictive Control for Dynamic Cloth Manipulation: Parameter Learning and Experimental Validation

  • Luque Acera, Adrià
  • Parent, David
  • Colomé, Adrià
  • Ocampo-Martínez, Carlos
  • Torras, Carme
Robotic cloth manipulation is a challenging problem for robotic systems. Textile items can adopt multiple configurations and shapes during their manipulation. Hence, robots should not only understand the current configuration of the item but also be able to predict its future possible behaviors and perform real-time control during manipulation. This article addresses the problem of indirectly controlling the configuration of certain points of a textile object, by applying actions on other parts of it through the use of a model predictive control (MPC) strategy. MPC allows to foresee the behavior of indirectly controlled points, while satisfying physical/operational constraints. This is done by first identifying the optimal control signals that may constitute the desired future cloth configuration. After that, a dynamic model of the item will be used and sensor data will allow to update the belief on the object's state and close the loop. This article investigates how grasping the upper corners of a square piece of cloth can allow to track a reference trajectory of the pieces' lower corners. To do so, we propose and validate a linear cloth model that allows solving the MPC optimization problem in real time. Reinforcement learning (RL) techniques are used to learn the parameters of the proposed cloth model and to tune the resulting MPC. The full control scheme was implemented and executed in a real robot, obtaining accurate tracking results in adverse conditions., This work was developed in the context of the project CLOTHILDE (”CLOTH manIpulation Learning from DEmonstrations”), which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programmee (Advanced Grant agreement No 741930). The work is also supported by Project PID2020- 118649RB-I00 funded by MCIN/ AEI /10.13039/501100011033, Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Social Assistive Robotics: An Ethical and Political Inquiry Through the Lens of Freedom

  • Pareto Boada, Júlia
  • Coeckelbergh, Mark
The development of social assistive robots for supporting healthcare provision faces a lack of an ethical approach that adequately addresses the normatively relevant challenges regarding its deployment. Current ethical reflection is primarily informed by an individual-centered perspective focused on robots' implications for their end-users and thereby limited to the dyadic human-robot interaction sphere. Considering that this is tightly correlated to the restricted understanding of core ethical concepts upon which reflection stands, this paper delves into the concept of freedom from a philosophical perspective to unfold its full normative breadth for a critical assessment of technological development. By bringing to the fore the political-structural dimension of freedom and, in turn, elaborating the political dimension of technology, the undertaken philosophical approach discloses freedom as a transversal ethical concept for a normative reflection on technology. Thereby, it broadens the scope of ethical attention beyond the sphere of human-robot interaction and turns attention to the so far overlooked structural dimension of human-robot relations. Drawing on conceptions of freedom as non-domination, among others, the paper approaches social assistive robotics and reexamines the terrain of relevant issues for its development. Since freedom is one major issue upon which current concerns revolve, the undertaken analysis substantially enriches the ongoing ethical discussion on social assistive robotics' implications for human freedom. In this way, this work contributes to going beyond the current individual-centered ethical perspective by laying conceptual grounds for a comprehensive ethical approach to social assistive robotics' development., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work has been partially supported by grant PRE2018-084286 funded by MCIN/ AEI /https://doi.org/10.13039/501100011033 and by "ESF Investing in your future", and by the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Learning in autonomous and intelligent systems: overview and biases from data sources, El aprendizaje en sistemas autónomos e inteligentes: visión general y sesgos de fuentes de datos

  • Jiménez Schlegl, Pablo
[EN] Autonomous and Intelligent Systems (A/IS, to adhere to the terminology of the IEEE Ethically Aligned Design report) can gather their knowledge by different means and from different sources. In principle, learning algorithms are neutral; rather, it is the data they are fed during the learning period that can introduce biases or a specific ethical orientation. Human control over the learning process is more straightforward in learning from demonstration, where data sources are restricted to the choices of the demonstrator (or teacher), but even in unsupervised versions of reinforcement learning, biases are present via the definition of the reward function. In this paper we provide an overview of learning paradigms of artificial systems: supervised and unsupervised methods, with the most striking examples in each category, without too much technical detail. Furthermore, we describe the types of data sources that are presently available and in use by the robotics community. We also focus on observable bias in image datasets and originated by human annotation. We point at quite recent research on bias in social robot navigation and end with a brief reflection about ambient influences on future learning robots., [ES] Los sistemas autónomos e inteligentes (A/IS por sus siglas en inglés, en concordancia con el informe del IEEE sobre diseño alineado con la ética) pueden obtener sus conocimientos a través de diferentes procedimientos y de fuentes diversas. Los algoritmos de aprendizaje son neutros en principio, son más bien los datos con los que se alimentan durante el período de aprendizaje que pueden introducir sesgos o una orientación ética específica. El control humano sobre el proceso de aprendizaje es más directo en aprendizaje por demostración, donde las fuentes de datos están restringidas a las elecciones del demostrador (o profesor), pero incluso en las versiones no supervisadas del aprendizaje por refuerzo los sesgos están presentes a través de la definición de la función de recompensa. En este artículo proporcionamos una visión general de los paradigmas de aprendizaje de los sistemas artificiales: métodos supervisados y no supervisados, con los ejemplos más destacados de cada categoría, sin profundizar demasiado en el detalle técnico. Además describimos los tipos de fuentes de datos disponibles actualmente y su uso por la comunidad robótica. También enfatizamos el sesgo que se observa en bases de datos de imágenes y originados por anotación humana. Destacamos una investigación muy reciente sobre sesgo en navegación de robots sociales y finalizamos con una breve reflexión sobre influencia del ambiente sobre futuros robots que aprenden., This work has been partially funded by the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE) and by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI[MDM-2016-0656] and the project HuMoUR TIN2017-90086-R., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Fitting Parameters of Linear Dynamical Systems to Regularize Forcing Terms in Dynamical Movement Primitives

  • Stulp, Freek
  • Colomé, Adrià
  • Torras, Carme
Due to their flexibility and ease of use, Dynamical Movement Primitives (DMPs) are widely used in robotics applications and research. DMPs combine linear dynamical systems to achieve robustness to perturbations and adaptation to moving targets with non-linear function approximators to fit a wide range of demonstrated trajectories.We propose a novel DMP formulation with a generalized logistic function as a delayed goal system. This formulation inherently has low initial jerk, and generates the bell-shaped velocity profiles that are typical of human movement. As the novel formulation is more expressive, it is able to fit a wide range of human demonstrations well, also without a non-linear forcing term. We exploit this increased expressiveness by automating the fitting of the dynamical system parameters through opti-mization. Our experimental evaluation demonstrates that this optimization regularizes the forcing term, and improves the interpolation accuracy of parametric DMPs., This work was partially funded by the European ERC/EIC/Horizon projects with grant numbers 741930 (CLOTHILDE), 951992 (VeriDream), and 101070596 (euROBIN), as well the CSIC project 202350E080 (ClothIRI), and the MCIN/AEI project PID2020-118649RBI00 (CHLOE-GRAPH)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

ProcSim: Proxy-based Confidence for Robust Similarity Learning

  • Barbany, Oriol
  • Lin, Xiaofan
  • Bastan, Muhammet
  • Dhua, Arnab
Deep Metric Learning (DML) methods aim at learning an embedding space in which distances are closely related to the inherent semantic similarity of the inputs. Previous studies have shown that popular benchmark datasets often contain numerous wrong labels, and DML methods are susceptible to them. Intending to study the effect of realistic noise, we create an ontology of the classes in a dataset and use it to simulate semantically coherent labeling mistakes. To train robust DML models, we propose ProcSim, a simple framework that assigns a confidence score to each sample using the normalized distance to its class representative. The experimental results show that the proposed method achieves state-of-the-art performance on the DML benchmark datasets injected with uniform and the proposed semantically coherent noise., The authors thank Amit Kumar K C, Michael Huang, and Rene Vidal for fruitful discussions and useful sugges- ´ tions. O.B. is part of CLOTHILDE (“CLOTH manIpulation Learning from DEmonstrations”) which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (Advanced Grant agreement No. 741930). O.B. thanks the European Laboratory for Learning and Intelligent Systems (ELLIS) for PhD program support., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Semantic State Estimation in Robot Cloth Manipulations Using Domain Adaptation from Human Demonstrations

  • Tzelepis, Georgies
  • Aksoy, Eren Erdal
  • Borràs, Julia
  • Alenyà, Guillem
Deformable object manipulations, such as those involving textiles, present a significant challenge due to their high dimensionality and complexity. In this paper, we propose a solution for estimating semantic states in cloth manipulation tasks. To this end, we introduce a new, large-scale, fully-annotated RGB image dataset of semantic states featuring a diverse range of human demonstrations of various complex cloth manipulations. This effectively transforms the problem of action recognition into a classification task. We then evaluate the generalizability of our approach by employing domain adaptation techniques to transfer knowledge from human demonstrations to two distinct robotic platforms: Kinova and UR robots. Additionally, we further improve performance by utilizing a semantic state graph learned from human manipulation data., This work has been partially funded by MCIN/ AEI /10.13039/ 501100011033, Spain, under the project CHLOE-GRAPH (PID2020- 119244GB-I00); by the European Union’s Horizon 2020 under ERC Advanced Grant CLOTHILDE (no. 741930); and by MCIN/ AEI /10.13039/501100011033, Spain, and by the “European Union NextGenerationEU/ PRTR, Spain, under the project COHERENT (PCI2020-120718-2);, Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Linear quadratic control of nonlinear systems with Koopman operator learning and the Nyström method

  • Caldarelli, Edoardo
  • Chatalic, Antoine
  • Colomé, Adrià
  • Molinari, Cesare
  • Ocampo-Martínez, Carlos
  • Torras, Carme
  • Rosasco, Lorenzo
In this paper, we study how the Koopman operator framework can be combined with kernel methods to effectively control nonlinear dynamical systems. While kernel methods have typically large computational requirements, we show how random subspaces (Nystr\"om approximation) can be used to achieve huge computational savings while preserving accuracy. Our main technical contribution is deriving theoretical guarantees on the effect of the Nystr\"om approximation. More precisely, we study the linear quadratic regulator problem, showing that the approximated Riccati operator converges at the rate $m^{-1/2}$, and the regulator objective, for the associated solution of the optimal control problem, converges at the rate $m^{-1}$, where $m$ is the random subspace size. Theoretical findings are complemented by numerical experiments corroborating our results., E. Caldarelli, A. Colomé and C. Torras acknowledge support from the project CLOTHILDE (“CLOTH manIpulation Learning from DEmonstrations”), funded by the European Research Council(ERC) under the European Union’s Horizon 2020 research and innovation programme (Advanced Grant agreement No 741930). E. Caldarelli acknowledges travel support from ELISE (GA No 951847). C. Molinari is part of “GNAMPA” (INdAM) and has been supported by the projects MIUR Excellence Department awarded to DIMA UniGe CUP D33C23001110001, AFOSR FA8655-22-1-7034, MIUR-PRIN 202244A7YL and PON “Ricerca e Innovazione” 2014–2020. C. Ocampo-Martinez acknowledges the support from the project SEAMLESS (PID2023-148840OB-I00) funded by MCIN/AEI. L. Rosasco acknowledges the financial support of the European Research Council, (grant SLING 819789), the European Commission, (Horizon Europe grant ELIAS 101120237), the US Air Force Office of Scientific Research (FA8655-22-1-7034), the Ministry of Education, University and Research (FARE grant ML4IP R205T7J2KP; grant BAC FAIR PE00000013 funded by the EU - NGEU) and the Center for Brains, Minds and Machines (CBMM) ., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Representació dels estats d’una tela basada en la derivada de la integral d’enllaç de Gauss

  • Coltraro, Franco
  • Fontana, Josep
  • Alberich-Carramiñana, Maria
  • Borràs, Julia
  • Torras, Carme
La manipulació robòtica de teles és una tasca complexa a causa de l’espai d’estats de forma d’infinites dimensions dels tèxtils, que dificulta molt la percepció dels estats. En aquest article introduïm les coordenades de tela dGLI, una representació de dimensió baixa dels estats d’una tela que ens permet distingir de manera eficient una gran varietat d’estats de plegament, fet que obre les portes a mètodes d’aprenentatge eficients per a la planificació i el control de la manipulació de teles. La representació es basa en una derivada direccional de la integral d’enllaç de Gauss i ens permet representar configuracions de plegament espacials i planàries de manera consistent i unificada. Es comprova que les coordenades de tela dGLI que proposem són més encertades en la representació dels estats de la tela i significativament més sensibles als canvis en possibilitats de prensió que altres mètodes clàssics de distància de formes. Per acabar, apliquem la representació a imatges reals d’una tela i es demostra que permet identificar els diferents estats utilitzant un classificador basat en la distància., Aquest estudi es va fer dins del projecte CLOTHILDE, que ha rebut finançament del Consell Europeu de Recerca (ERC) dins del programa de recerca i innovació de la UE Horitzó 2020 (acord de subvenció núm. 741930). M. Alberich-Carramiñana també està afiliada a la BarcelonaGraduate School of Mathematics (BGSMath) i a l’Institut de Matemàtiques de la UPC-BarcelonaTech (IMTech), i tant ella com J. Amorós estan finançats parcialment per l’Agència Estatal d’Investigació espanyola AEI/10.13039/501100011033 subvenció PID2019-103849GB-I00 i pel pro-jecte AGAUR 2021 SGR 00603 Geometry of Manifolds and Applications, GEOMVAP. J. Borràs està finançada per l’Agència Estatal d’Investigació espanyola MCIN/AEI/10.13039/501100011033 subvenció PID2020-118649RB-I00 (projecte CHLOE-GRAPH), Peer reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

The effects of gender and personality of robot assistants on customers’ acceptance of their service

  • Forgas Coll, Santiago
  • Huertas Garcia, Rubén
  • Andriella, Antonio|||0000-0002-6641-6450
  • Alenyà Ribas, Guillem|||0000-0002-6018-154X
The Covid-19 pandemic has stimulated the use of social robots in front-office services. However, some initial applications yielded disappointing results, as managers were unaware of the level of development of the robots’ artificial intelligence systems. This study proposes to adapt the Almere model to estimate the technological acceptance of service robots, which express their gender and personality, whilst assisting consumers. A 2¿×¿2 (two genders vs. two personalities) between-subjects experiment was conducted with 219 participants. Model estimation with Structural Equation Modelling confirmed seven out of eight hypotheses, and all four scenarios were estimated with Ordinary Least Squares, showing that robot gender and personality affected their technological acceptance., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Ethics of social robotics: individual and societal concerns and opportunities

  • Torras, Carme|||0000-0002-2933-398X
© 2024 by the author(s). This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. See credit lines of images or other third-party material in this article for license information., Focus on the ethics of a given technology tends to lag far behind its development. This lag has been particularly acute in the case of artificial intelligence, whose accelerated deployment in a wide range of domains has triggered unprecedented attention on the risks and consequences for society at large, leading to a myriad of ethics regulations, which are difficult to coordinate and integrate due to their late appearance. The very nature of social robots forces their deployment to occur at a much slower pace, providing an opportunity for a profound reflection on ethics, which is already happening in multidisciplinary teams. This article provides a personal view of the ethics landscape, centered on the particularities of social robotics, with the main issues being ordered along two axes (individual and societal) and grouped into eight categories (human dignity, human autonomy, robot transparency, emotional bonding, privacy and safety, justice, freedom, and responsibility). This structure stems from the experience of developing and teaching a university course on ethics in social robotics, whose pedagogical materials are freely available., This work was partly supported by the European Research Council through the project CLOTHILDE (Cloth Manipulation Learning from Demonstrations; advanced grant agreement 741930) and by the European Union's Horizon Europe program through the project SoftEnable (HORIZON-CL4-2021-DIGITAL-EMERGING-01-101070600)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Generating predicate suggestions based on the space of plans: an example of planning with preferences

  • Canal Camprodon, Gerard|||0000-0002-6718-1198
  • Torras, Carme|||0000-0002-2933-398X
  • Alenyà Ribas, Guillem|||0000-0002-6018-154X
The version of record is available online at: https://dx.doi.org/10.1007/s11257-022-09327-w, Task planning in human–robot environments tends to be particularly complex as it involves additional uncertainty introduced by the human user. Several plans, entailing few or various differences, can be obtained to solve the same given task. To choose among them, the usual least-cost plan criteria is not necessarily the best option, because here, human constraints and preferences come into play. Knowing these user preferences is very valuable to select an appropriate plan, but the preference values are usually hard to obtain. In this context, we propose the Space-of-Plans-based Suggestions (SoPS) algorithms that can provide suggestions for some planning predicates, which are used to define the state of the environment in a task planning problem where actions modify the predicates. We denote these predicates as suggestible predicates, of which user preferences are a particular case. The first algorithm is able to analyze the potential effect of the unknown predicates and provide suggestions to values for these unknown predicates that may produce better plans. The second algorithm is able to suggest changes to already known values that potentially improve the obtained reward. The proposed approach utilizes a Space of Plans Tree structure to represent a subset of the space of plans. The tree is traversed to find the predicates and the values that would most increase the reward, and output them as a suggestion to the user. Our evaluation in three preference-based assistive robotics domains shows how the proposed algorithms can improve task performance by suggesting the most effective predicate values first., This work has been partially supported by the ERC project Clothilde (ERC-2016-ADG-741930); by MCIN/AEI/10.13039/501100011033 under the project CHLOE-GRAPH (PID2020-118649RB-l00); by the European Union NextGenerationEU/PRTR under the project ROB-IN (PLEC2021-007859); and by the CHIST-ERA project COHERENT (EPSRC EP/V062506/1), Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Exploiting symmetries in reinforcement learning of bimanual robotic tasks

  • Amadio, Fabio
  • Colomé Figueras, Adrià|||0000-0001-9715-4062
  • Torras, Carme|||0000-0002-2933-398X
© 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Movement Primitives (MPs) have been widely adopted for representing and learning robotic movements using Reinforcement Learning Policy Search. Probabilistic Movement Primitives (ProMPs) are a kind of MP based on a stochastic representation over sets of trajectories, able of capturing the variability allowed while executing a movement. This approach has proved effective in learning a wide range of robotic movements, but it comes with the need of dealing with a high-dimensional space of parameters. This may be a critical problem when learning tasks with two robotic manipulators, and this work proposes an approach to reduce the dimension of the parameter space based on the exploitation of symmetry. A symmetrization method for ProMPs is presented and used to represent two movements, employing a single ProMP for the first arm and a symmetry surface that maps that ProMP to the second arm. This symmetric representation is then adopted in reinforcement learning of bimanual tasks (from user-provided demonstrations), using Relative Entropy Policy Search (REPS) algorithm. The symmetry-based approach developed has been tested in an experiment of cloth manipulation, showing a speed increment in learning the task., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Benchmarking the sim-to-real gap in cloth manipulation

  • Blanco Mulero, David|||0000-0003-1866-2144
  • Barbany Mayor, Oriol|||0000-0002-1379-9152
  • Alcan, Gokhan
  • Colomé Figueras, Adrià|||0000-0001-9715-4062
  • Torras, Carme|||0000-0002-2933-398X
  • Kyrki, Ville
Realistic physics engines play a crucial role for learning to manipulate deformable objects such as garments in simulation. By doing so, researchers can circumvent challenges such as sensing the deformation of the object in the real-world. In spite of the extensive use of simulations for this task, few works have evaluated the reality gap between deformable object simulators and real-world data. We present a benchmark dataset to evaluate the sim-to-real gap in cloth manipulation. The dataset is collected by performing a dynamic as well as a quasi-static cloth manipulation task involving contact with a rigid table. We use the dataset to evaluate the reality gap, computational time, and simulation stability of four popular deformable object simulators: MuJoCo, Bullet, Flex, and SOFA. Additionally, we discuss the benefits and drawbacks of each simulator. The benchmark dataset is open-source. Supplementary material, videos, and code, can be found at https://sites.google.com/view/cloth-sim2real-benchmark., This work was supported in part by Academy of Finland under Grant 317020, in part by Business Finland (decision 9249/31/2021), in part by EU’s H2020 Programme through the CLOTHILDE Project under Grant ERC-2016-ADG-741930, in part by EU’s Horizon Europe Programme through the SoftEnable Project under Grant Agreement 101070600, and in part by MCIN/AEI/10.13039/501100011033 through CHLOE-GRAPH Project under Grant PID2020-118649RB-I00., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Social robot-delivered customer-facing services: an assessment of the experience

  • Forgas Coll, Santiago
  • Huertas Garcia, Rubén
  • Andriella, Antonio|||0000-0002-6641-6450
  • Alenyà Ribas, Guillem|||0000-0002-6018-154X
The ability to install social intelligence protocols in robots in order for them to exhibit conversational skills has made them ideal tools for delivering services with a high cognitive and low emotional load. Little is known about how this capability influences the customer experience and the intention to continue receiving these services. Experiences were assessed in a study simulating customer-facing service delivery, and the constructs of the technology readiness index and stated gender were analysed as possible moderators in a quasi-experiment. Hedonic quality was the most relevant factor explaining attitude, and attitude explained intention to use as well as social influence. As for the constructs of technological readiness and gender, optimism and innovativeness seem to be the most likely candidates for moderating the other variables. The most optimistic and the most innovative route would be for the main actors to continue adapting to social robot technology in the future., This work was partially funded by the European Union’s Horizon 2020 under ERC Advanced Grant CLOTHILDE (no. 741930); by MCIN/ AEI /10.13039/501100011033 and by the ”European Union NextGenerationEU/PRTR” under the project ROB-IN (PLEC2021-007859); by the Research Council of Norway under the project SECUROPS (INT-NO/0875); and by the ”European Union NextGenerationEU/PRTR” through CSIC’s Thematic Platforms (PTI+ Neuro-Aging)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Are preferences useful for better assistance? A physically assistive robotics user study

  • Canal Camprodon, Gerard|||0000-0002-6718-1198
  • Torras, Carme|||0000-0002-2933-398X
  • Alenyà Ribas, Guillem|||0000-0002-6018-154X
© 2021 Copyright held by the owner/author(s)., Assistive Robots have an inherent need of adapting to the user they are assisting. This is crucial for the correct development of the task, user safety, and comfort. However, adaptation can be performed in several manners. We believe user preferences are key to this adaptation. In this paper, we evaluate the use of preferences for Physically Assistive Robotics tasks in a Human-Robot Interaction user evaluation. Three assistive tasks have been implemented consisting of assisted feeding, shoe-fitting, and jacket dressing, where the robot performs each task in a different manner based on user preferences. We assess the ability of the users to determine which execution of the task used their chosen preferences (if any). The obtained results show that most of the users were able to successfully guess the cases where their preferences were used even when they had not seen the task before. We also observe that their satisfaction with the task increases when the chosen preferences are employed. Finally, we also analyze the user’s opinions regarding assistive tasks and preferences, showing promising expectations as to the benefits of adapting the robot behavior to the user through preferences., This work has been supported by the ERC project Clothilde (ERC-2016-ADG-741930), the HuMoUR project (Spanish Ministry of Science and Innovation TIN2017-90086-R) and by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656). Gerard Canal has also been supported by the Spanish Ministry of Education, Culture and Sport by the FPU15/00504 doctoral grant and the CHIST-ERA project COHERENT (EPSRC EP/V062506/1)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Learning in autonomous and intelligent systems: Overview and biases from data sources

  • Jiménez Schlegl, Pablo|||0000-0003-3627-4938
© 2021 CSIC. Este es un artículo de acceso abierto distribuido bajo los términos de la licencia de uso y distribu-ción Creative Commons Reconocimiento 4.0 Internacional (CC BY 4.0)., Autonomous and Intelligent Systems (A/IS, to adhere to the terminology of the IEEE Ethically Aligned Design report) can gather their knowledge by different means and from different sources. In principle, learning algorithms are neutral; rather, it is the data they are fed during the learning period that can introduce biases or a specific ethical orientation. Human control over the learning process is more straightforward in learning from demonstration, where data sources are restricted to the choices of the demonstrator (or teacher), but even in unsupervised versions of reinforcement learning, biases are present via the definition of the reward function. In this paper we provide an overview of learning paradigms of artificial systems: supervised and unsupervised methods, with the most striking examples in each category, without too much technical detail. Furthermore, we describe the types of data sources that are presently available and in use by the robotics community. We also focus on observable bias in image datasets and originated by human annotation. We point at quite recent research on bias in social robot navigation and end with a brief reflection about ambient influences on future learning robots., This work has been partially funded by the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE) and by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI[MDM-2016-0656] and the project HuMoUR TIN2017-90086-R., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Do I have a personality? Endowing care robots with context-dependent personality traits

  • Andriella, Antonio|||0000-0002-6641-6450
  • Siqueira, Henrique
  • Fu, Di
  • Magg, Sven
  • Barros, Pablo
  • Wermter, Stefan
  • Torras, Carme|||0000-0002-2933-398X
  • Alenyà Ribas, Guillem|||0000-0002-6018-154X
Recent studies have revealed the key importance of modelling personality in robots to improve interaction quality by empowering them with social-intelligence capabilities. Most research relies on verbal and non-verbal features related to personality traits that are highly context-dependent. Hence, analysing how humans behave in a given context is crucial to evaluate which of those social cues are effective. For this purpose, we designed an assistive memory game, in which participants were asked to play the game obtaining support from an introvert or extroverted helper, whether from a human or robot. In this context, we aim to (i) explore whether selective verbal and non-verbal social cues related to personality can be modelled in a robot, (ii) evaluate the efficiency of a statistical decision-making algorithm employed by the robot to provide adaptive assistance, and (iii) assess the validity of the similarity attraction principle. Specifically, we conducted two user studies. In the human–human study (N=31), we explored the effects of helper’s personality on participants’ performance and extracted distinctive verbal and non-verbal social cues from the human helper. In the human–robot study (N=24), we modelled the extracted social cues in the robot and evaluated its effectiveness on participants’ performance. Our findings showed that participants were able to distinguish between robots’ personalities, and not between the level of autonomy of the robot (Wizard-of-Oz vs fully autonomous). Finally, we found that participants achieved better performance with a robot helper that had a similar personality to them, or a human helper that had a different personality., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Social networks and robot companions: technology, ethics and science fiction

  • Torras, Carme|||0000-0002-2933-398X
Information technologies have become part of our everyday lives and are increasingly acting as intermediaries in our workplaces and personal relationships or even substituting them. This growing interaction with machines poses several questions about which we have no previous experience, nor can we reliably predict how they will influence the evolution of society. This has led to the convergence of technoscience and humanities in an ethical debate that is starting to bear fruit, not only with the setting of regulations and standards, but also with educational initiatives in university teaching, professional improvement, and the conformation of public opinion. Interestingly, science fiction often plays a prominent speculative role in highlighting the pros and cons of potential scenarios., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Task-adaptive robot learning from demonstration with gaussian process models under replication

  • Arduengo García, Miguel
  • Colomé Figueras, Adrià|||0000-0001-9715-4062
  • Borràs Sol, Júlia|||0000-0002-1662-2037
  • Sentís Álvarez, Luis
  • Torras, Carme|||0000-0002-2933-398X
© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Learning from Demonstration (LfD) is a paradigm that allows robots to learn complex manipulation tasks that can not be easily scripted, but can be demonstrated by a human teacher. One of the challenges of LfD is to enable robots to acquire skills that can be adapted to different scenarios. In this paper, we propose to achieve this by exploiting the variations in the demonstrations to retrieve an adaptive and robust policy, using Gaussian Process (GP) models. Adaptability is enhanced by incorporating task parameters into the model, which encode different specifications within the same task. With our formulation, these parameters can be either real, integer, or categorical. Furthermore, we propose a GP design that exploits the structure of replications, i.e., repeated demonstrations with identical conditions within data. Our method significantly reduces the computational cost of model fitting in complex tasks, where replications are essential to obtain a robust model. We illustrate our approach through several experiments on a handwritten letter demonstration dataset., This work has been partially funded by the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE) and by the Spanish State Research Agency through the Mar ́ıa de Maeztu Seal of Excellence to IRI [MDM-2016-0656]., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Model predictive control for dynamic cloth manipulation: parameter learning and experimental validation

  • Luque Acera, Adrià
  • Parent Alonso, David
  • Colomé Figueras, Adrià|||0000-0001-9715-4062
  • Ocampo-Martínez, Carlos|||0000-0001-9251-6044
  • Torras, Carme|||0000-0002-2933-398X
© 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Robotic cloth manipulation is a challenging problem for robotic systems. Textile items can adopt multiple configurations and shapes during their manipulation. Hence, robots should not only understand the current configuration of the item but also be able to predict its future possible behaviours and perform real-time control during manipulation. This paper addresses the problem of indirectly controlling the configuration of certain points of a textile object, by applying actions on other parts of it through the use of a Model Predictive Control (MPC) strategy. MPC allows to foresee the behavior of indirectly controlled points, while satisfying physical/operational constraints. This is done by first identifying the optimal control signals that may constitute the desired future cloth configuration. After that, a dynamic model of the item will be used and sensor data will allow to update the belief on the object’s state and close the loop. This paper investigates how grasping the upper corners of a square piece of cloth can allow to track a reference trajectory of the pieces’ lower corners. To do so, we propose and validate a linear cloth model that allows solving the MPC optimization problem in real time. Reinforcement Learning (RL) techniques are used to learn the parameters of the proposed cloth model and to tune the resulting MPC. The full control scheme was implemented and executed in a real robot, obtaining accurate tracking results in adverse conditions., This work was developed in the context of the project CLOTHILDE (”CLOTH manIpulation Learning from DEmonstrations”), which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programmee (Advanced Grant agreement No 741930). The work is also supported by Project PID2020- 118649RB-I00 funded by MCIN/ AEI /10.13039/501100011033., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Dimensionality reduction in learning Gaussian mixture models of movement primitives for contextualized action selection and adaptation

  • Colomé Figueras, Adrià|||0000-0001-9715-4062
  • Torras, Carme|||0000-0002-2933-398X
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Robotic manipulation often requires adaptation to changing environments. Such changes can be represented by a certain number of contextual variables that may be observed or sensed in different manners. When learning and representing robot motion –usually with movement primitives–, it is desirable to adapt the learned behaviors to the current context. Moreover, different actions or motions can be considered in the same framework, using contextualization to decide which action applies to which situation. Such frameworks, however, may easily become large-dimensional, thus requiring to reduce the dimensionality of the parameters space, as well as the amount of data needed to generate and improve the model over experience. In this paper, we propose an approach to obtain a generative model from a set of actions that share a common feature. Such feature, namely a contextual variable, is plugged into the model to generate motion. We encode the data with a Gaussian Mixture Model in the parameter space of Probabilistic Movement Primitives (ProMPs), after performing Dimensionality Reduction (DR) on such parameter space. We append the contextual variable to the parameter space and obtain the number of Gaussian components, i.e., different actions in a dataset, through Persistent Homology. Then, using multimodal Gaussian Mixture Regression (GMR), we can retrieve the most likely actions given a contextual situation and execute them. After actions are executed, we use Reward-Weighted Responsibility GMM (RWR-GMM) update the model after each execution. Experimentation in 3 scenarios shows that the method drastically reduces the dimensionality of the parameter space, thus implementing both action selection and adaptation to a changing situation in an efficient way., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

A representation of cloth states based on a derivative of the Gauss linking integral

  • Coltraro, Franco|||0000-0002-9149-950X
  • Fontana, Josep
  • Amorós Torrent, Jaume|||0000-0002-3133-9985
  • Alberich Carramiñana, Maria|||0000-0003-2749-4875
  • Borràs Sol, Júlia|||0000-0002-1662-2037
  • Torras, Carme|||0000-0002-2933-398X
Robotic manipulation of cloth is a complex task because of the infinite-dimensional shape-state space of textiles, which makes their state estimation very difficult. In this paper we introduce the dGLI Cloth Coordinates, a finite low-dimensional representation of cloth states that allows us to efficiently distinguish a large variety of different folded states, opening the door to efficient learning methods for cloth manipulation planning and control. Our representation is based on a directional derivative of the Gauss Linking Integral and allows us to represent spatial as well as planar folded configurations in a consistent and unified way. The proposed dGLI Cloth Coordinates are shown to be more accurate in the representation of cloth states and significantly more sensitive to changes in grasping affordances than other classic shape distance methods. Finally, we apply our representation to real images of a cloth, showing that with it we can identify the different states using a distance-based classifier., This work was developed under the project CLOTHILDE which has received funding from the European Research Council (ERC) under the EU-Horizon 2020 research and innovation programme (grant agreement No. 741930). M. Alberich-Carramiñana is also with the Barcelona Graduate School of Mathematics (BGSMath) and the Institut de Matemàtiques de la UPC-BarcelonaTech (IMTech), and she and J. Amorós are partially supported by the Spanish State Research Agency AEI/10.13039/501100011033 grant PID2019-103849GB-I00 and by the AGAUR project 2021 SGR 00603 Geometry of Manifolds and Applications, GEOMVAP. J. Borràs is supported by the Spanish State Research Agency MCIN/ AEI /10.13039/501100011033 grant PID2020-118649RB-I00 (CHLOE-GRAPH project)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Social assistive robotics: an ethical and political inquiry through the lens of freedom

  • Pareto Boada, Júlia|||0000-0003-4879-8800
  • Coeckelbergh, Mark
The development of social assistive robots for supporting healthcare provision faces a lack of an ethical approach that adequately addresses the normatively relevant challenges regarding its deployment. Current ethical reflection is primarily informed by an individual-centered perspective focused on robots’ implications for their end-users and thereby limited to the dyadic human–robot interaction sphere. Considering that this is tightly correlated to the restricted understanding of core ethical concepts upon which reflection stands, this paper delves into the concept of freedom from a philosophical perspective to unfold its full normative breadth for a critical assessment of technological development. By bringing to the fore the political-structural dimension of freedom and, in turn, elaborating the political dimension of technology, the undertaken philosophical approach discloses freedom as a transversal ethical concept for a normative reflection on technology. Thereby, it broadens the scope of ethical attention beyond the sphere of human–robot interaction and turns attention to the so far overlooked structural dimension of human–robot relations. Drawing on conceptions of freedom as non-domination, among others, the paper approaches social assistive robotics and reexamines the terrain of relevant issues for its development. Since freedom is one major issue upon which current concerns revolve, the undertaken analysis substantially enriches the ongoing ethical discussion on social assistive robotics’ implications for human freedom. In this way, this work contributes to going beyond the current individual-centered ethical perspective by laying conceptual grounds for a comprehensive ethical approach to social assistive robotics’ development., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work has been partially supported by grant PRE2018-084286 funded by MCIN/ AEI /https://doi.org/10.13039/501100011033 and by "ESF Investing in your future", and by the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

A Versatile Gripper for Cloth Manipulation

  • Donaire Cónsul, Sònia
  • Borràs Sol, Júlia|||0000-0002-1662-2037
  • Alenyà Ribas, Guillem|||0000-0002-6018-154X
  • Torras, Carme|||0000-0002-2933-398X
Cloth manipulation has been mostly advancing in perception and modeling methods for cloth state estimation and grasping point detection. In comparison, less attention has been put on end-effector design. Indeed, most implementations use a parallel gripper that can only perform pinch grasps. Instead, a more versatile set of possible grasp types could ease many tasks by providing more support to certain parts of the object, as well as make feasible tasks that become very complex when executed with only pinch grasps. We present a versatile gripper design which, besides the common open-close thumb+finger(s) feature, it has a couple of reconfiguration degrees of freedom that offer, first, a wide base plane to provide support and, second, variable friction surface on the thumb tip. Our gripper can execute a versatile set of grasps that ease some complex tasks such as pick and place folded clothes or fold in the air. In addition, the variable friction mechanism enables a more robust pinch-and-slide manipulation to trace cloth edges. Our evaluation shows the gripper potential to execute a wide variety of cloth manipulation tasks., This work receives funding from the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE), Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656) and the “Ramon y Cajal” Fellowship RYC-2017-22703
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Benchmarking bimanual cloth manipulation

  • Garcia Camacho, Irene|||0000-0002-1210-7345
  • Lippi, Martina
  • Welle, Michael C.
  • Yin, Hang
  • Antonova, Rika
  • Varava, Anastasiia
  • Borràs Sol, Júlia|||0000-0002-1662-2037
  • Torras, Carme|||0000-0002-2933-398X
  • Marino, Alessandro
  • Alenyà Ribas, Guillem|||0000-0002-6018-154X
  • Kragic, Danica
© 2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Cloth manipulation is a challenging task that, despite its importance, has received relatively little attention compared to rigid object manipulation. In this paper, we provide three benchmarks for evaluation and comparison of different approaches towards three basic tasks in cloth manipulation: spreading a tablecloth over a table, folding a towel, and dressing. The tasks can be executed on any bimanual robotic platform and the objects involved in the tasks are standardized and easy to acquire. We provide several complexity levels for each task, and describe the quality measures to evaluate task execution. Furthermore, we provide baseline solutions for all the tasks and evaluate them according to the proposed metrics., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

A Grasping-centered analysis for cloth manipulation

  • Borràs Sol, Júlia|||0000-0002-1662-2037
  • Alenyà Ribas, Guillem|||0000-0002-6018-154X
  • Torras, Carme|||0000-0002-2933-398X
© 2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Compliant and soft hands have gained a lot of atten- tion in the past decade because of their ability to adapt to the shape of the objects, increasing their effectiveness for grasping. However, when it comes to grasping highly flexible objects such as textiles, we face the dual problem: it is the object that will adapt to the shape of the hand or gripper. In this context, the classic grasp analysis or grasping taxonomies are not suitable for describing textile objects grasps. This article proposes a novel definition of textile object grasps that abstracts from the robotic embodiment or hand shape and recovers concepts from the early neuroscience literature on hand prehension skills. This framework enables us to identify what grasps have been used in literature until now to perform robotic cloth manipulation, and allows for a precise definition of all the tasks that have been tackled in terms of manipulation primitives based on regrasps. In addition, we also review what grippers have been used. Our analysis shows how the vast majority of cloth manipulations have relied only on one type of grasp, and at the same time we identify several tasks that need more variety of grasp types to be executed successfully. Our framework is generic, provides a classification of cloth manipulation primitives and can inspire gripper design and benchmark construction for cloth manipulation., This workwas supported in part by the European Research Council from the EuropeanUnion Horizon 2020 Programme under Grant 741930 (CLOTHILDE: CLOTHmanIpulation Learning from DEmonstrations), in part by the Spanish StateResearch Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656), and in part by the “Ramon y Cajal” Fellowship RYC-2017-22703., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Gendered human–robot Interactions in services

  • Forgas Coll, Santiago
  • Huertas Garcia, Rubén
  • Andriella, Antonio|||0000-0002-6641-6450
  • Alenyà Ribas, Guillem|||0000-0002-6018-154X
© The Author(s) 2023, The outbreak of Covid-19 precipitated the use of service robots in customer-facing services as a replacement for employ- ees to avoid human-to-human contact. However, this development has not resolved the debate as to whether robots should be characterized with gender attributes or simply be genderless. This study explores whether endowing a robot with gender attributes makes it more acceptable as a service provider among stated men and women. To this end, an experiment was conducted at a public fair in which a gendered robot simulated the provision of a service to customers, which consisted of offering them advice, hints, and messages of encouragement to help complete a eudaemonic puzzle. A parsimonious version of the Almere model was used to estimate acceptance of the technology. The findings reveal that for both stated men and women, the main drivers for accepting the female-coded robot are perceived usefulness and social influence, although women attach greater importance to social influence. For the male-coded robot, perceived usefulness and social influence are the main arguments for women, while for men they are enjoyment, perceived usefulness and, negatively, ease of use. In addition, different indirect effects between stated sexes are also identified. In summary, men and women consider different factors when accepting robots of each gender., This work was partially funded by the European Union’s Horizon 2020 programme under ERC Advanced Grant CLOTHILDE (no. 741930) and under the Marie Skłodowska-Curie grant agreement (no. 712949) (TECNIOspring PLUS); by MCIN/ AEI /https://doi.org/10.13039/501100011033 and by the “European Union NextGenerationEU/PRTR” under the project ROB-IN (PLEC2021-007859); by the Research Council of Norway under the project SECUROPS (INT-NO/0875); and by the “European Union NextGenerationEU/PRTR” through CSIC’s Thematic Platforms (PTI + Neuro-Aging)., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Gaussian-process-based robot learning from demonstration

  • Arduengo García, Miguel
  • Colomé Figueras, Adrià|||0000-0001-9715-4062
  • Lobo Prat, Joan
  • Sentis, Luis
  • Torras, Carme|||0000-0002-2933-398X
© The Author(s) 2023. Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This article is licensed under a Creative Commons Attri- bution 4.0 International License,, Learning from demonstration allows to encode task constraints from observing the motion executed by a human teacher. We present a Gaussian-process-based learning from demonstration (LfD) approach that allows robots to learn manipulation skills from demonstrations of a human teacher. By exploiting the potential that Gaussian process (GP) models offer, we unify in a single, entirely GP-based framework, the main features required for a state-of-the-art LfD approach. We address how GP can be used to effectively learn a policy from trajectories in task space. To achieve an effective generalization across demonstrations, we propose the novel Task Completion Index (TCI) for temporal alignment of task trajectories. Also, our probabilistic GP-based representation allows encoding variability throughout the different phases of the task. Finally, we present a method to efficiently adapt the policy to fulfill new requirements and modulate the robot behavior as a function of task variability. This approach has been successfully tested in a real-world application, namely teaching a TIAGo robot to open different types of doors., This work is partially funded by ERC Advanced Grant H2020-741930 (project CLOTHILDE), Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Controlled Gaussian process dynamical models with application to robotic cloth manipulation

  • Amadio, Fabio
  • Delgado Guerrero, Juan Antonio
  • Colomé Figueras, Adrià|||0000-0001-9715-4062
  • Torras, Carme|||0000-0002-2933-398X
Over the last years, significant advances have been made in robotic manipulation, but still, the handling of non-rigid objects, such as cloth garments, is an open problem. Physical interaction with non-rigid objects is uncertain and complex to model. Thus, extracting useful information from sample data can considerably improve modeling performance. However, the training of such models is a challenging task due to the high-dimensionality of the state representation. In this paper, we propose Controlled Gaussian Process Dynamical Models (CGPDMs) for learning high-dimensional, nonlinear dynamics by embedding them in a low-dimensional manifold. A CGPDM is constituted by a low-dimensional latent space, with an associated dynamics where external control variables can act and a mapping to the observation space. The parameters of both maps are marginalized out by considering Gaussian Process priors. Hence, a CGPDM projects a high-dimensional state space into a smaller dimension latent space, in which it is feasible to learn the system dynamics from training data. The modeling capacity of CGPDM has been tested in both a simulated and a real scenario, where it proved to be capable of generalizing over a wide range of movements and confidently predicting the cloth motions obtained by previously unseen sequences of control actions., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work was partially developed in the context of the project CLOTHILDE (“CLOTH manIpulation Learning from DEmonstrations”), which has received funding from ERC under the European Union’s Horizon 2020 research and innovation program (Advanced Grant agreement No 741930). This work has also received funding from project CHLOE-GRAPH (PID2020-118649RB-I00) funded by MCIN/ AEI /10.13039/501100011033., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Perception of cloth in assistive robotic manipulation tasks

  • Jiménez Schlegl, Pablo|||0000-0003-3627-4938
  • Torras, Carme|||0000-0002-2933-398X
The final publication is available at link.springer.com, Assistive robots need to be able to perform a large number of tasks that imply some type of cloth manipulation. These tasks include domestic chores such as laundry handling or bed-making, among others, as well as dressing assistance to disabled users. Due to the deformable nature of fabrics, this manipulation requires a strong perceptual feedback. Common perceptual skills that enable robots to complete their cloth manipulation tasks are reviewed here, mainly relying on vision, but also resorting to touch and force. The use of such basic skills is then examined in the context of the different cloth manipulation tasks, be them garment-only applications in the line of performing domestic chores, or involving physical contact with a human as in dressing assistance., This work was developed in the context of the project CLOTHILDE (“CLOTH manIpulation Learning from DEmonstra-tions”), which has received funding from the European Research Council (ERC) under the European Union’s Horizon2020research and innovation programme (Advanced Grant agreement No 741930). This work is supported by the Spanish StateResearch Agency through the Mar ́ıa de Maeztu Seal of Excellence to IRI (MDM-2016-065, Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Dimensionality reduction for dynamic movement primitives and application to bimanual manipulation of clothes

  • Colomé Figueras, Adrià|||0000-0001-9715-4062
  • Torras, Carme|||0000-0002-2933-398X
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Dynamic Movement Primitives (DMPs) are nowadays widely used as movement parametrization for learning robot trajectories, because of their linearity in the parameters, rescaling robustness and continuity. However, when learning a movement with DMPs, a very large number of Gaussian approximations needs to be performed. Adding them up for all joints yields too many parameters to be explored when using Reinforcement Learning (RL), thus requiring a prohibitive number of experiments/simulations to converge to a solution with a (locally or globally) optimal reward. In this paper we address the process of simultaneously learning a DMP-characterized robot motion and its underlying joint couplings through linear Dimensionality Reduction (DR), which will provide valuable qualitative information leading to a reduced and intuitive algebraic description of such motion. The results in the experimental section show that not only can we effectively perform DR on DMPs while learning, but we can also obtain better learning curves, as well as additional information about each motion: linear mappings relating joint values and some latent variables., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Linear quadratic control of nonlinear systems with Koopman operator learning and the Nyström method

  • Caldarelli, Edoardo
  • Chatalic, Antoine
  • Colomé Figueras, Adrià|||0000-0001-9715-4062
  • Molinari, Cesare
  • Ocampo-Martínez, Carlos|||0000-0001-9251-6044
  • Torras, Carme|||0000-0002-2933-398X
  • Rosasco, Lorenzo
© 2025 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/), In this paper, we study how the Koopman operator framework can be combined with kernel methods to effectively control nonlinear dynamical systems. While kernel methods have typically large computational requirements, we show how random subspaces (Nyström approximation) can be used to achieve huge computational savings while preserving accuracy. Our main technical contribution is deriving theoretical guarantees on the effect of the Nyström approximation. More precisely, we study the linear quadratic regulator problem, showing that the approximated Riccati operator converges at the rate m^(-1/2), and the regulator objective, for the associated solution of the optimal control problem, converges at the rate m^-1, where m is the random subspace size. Theoretical findings are complemented by numerical experiments corroborating our results., E. Caldarelli, A. Colomé and C. Torras acknowledge support from the project CLOTHILDE (“CLOTH manIpulation Learning from DEmonstrations”), funded by the European Research Council(ERC) under the European Union’s Horizon 2020 research and innovation programme (Advanced Grant agreement No 741930). E. Caldarelli acknowledges travel support from ELISE (GA No 951847). C. Molinari is part of “GNAMPA” (INdAM) and has been supported by the projects MIUR Excellence Department awarded to DIMA UniGe CUP D33C23001110001, AFOSR FA8655-22-1-7034, MIUR-PRIN 202244A7YL and PON “Ricerca e Innovazione” 2014–2020. C. Ocampo-Martinez acknowledges the support from the project SEAMLESS (PID2023-148840OB-I00) funded by MCIN/AEI. L. Rosasco acknowledges the financial support of the European Research Council, (grant SLING 819789), the European Commission, (Horizon Europe grant ELIAS 101120237), the US Air Force Office of Scientific Research (FA8655-22-1-7034), the Ministry of Education, University and Research (FARE grant ML4IP R205T7J2KP; grant BAC FAIR PE00000013 funded by the EU - NGEU) and the Center for Brains, Minds and Machines (CBMM) ., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Deformable surface reconstruction via Riemannian metric preservation

  • Barbany Mayor, Oriol|||0000-0002-1379-9152
  • Colomé Figueras, Adrià|||0000-0001-9715-4062
  • Torras, Carme|||0000-0002-2933-398X
Estimating the pose of an object from a monocular image is a fundamental inverse problem in computer vision. Due to its ill-posed nature, solving this problem requires incorporating deformation priors. In practice, many materials do not perceptibly shrink or extend when manipulated, constituting a reliable and well-known prior. Mathematically, this translates to the preservation of the Riemannian metric. Neural networks offer the perfect playground to solve the surface reconstruction problem as they can approximate surfaces with arbitrary precision and allow the computation of differential geometry quantities. This paper presents an approach for inferring continuous deformable surfaces from a sequence of images, which is benchmarked against several techniques and achieves state-of-the-art performance without the need for offline training. Being a method that performs per-frame optimization, our method can refine its estimates, contrary to those based on performing a single inference step. Despite enforcing differential geometry constraints at each update, our approach is the fastest of all the tested optimization-based methods., This work is part of the project CLOTHILDE (‘‘CLOTH manIpulation Learning from DEmonstrations’’) which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (Advanced Grant agreement No. 741930)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Topological representation of cloth state for robot manipulation: deriving the configuration space of a rectangular cloth

  • Strazzeri, Fabio
  • Torras, Carme|||0000-0002-2933-398X
The version of record is available online at: https://doi.org/10.1007/s10514-021-09968-7, Forty years ago the notion of configuration space (C-space) revolutionised robot motion planning for rigid and articulated objects. Despite great progress, handling deformable materials has remained elusive because of their infinite-dimensional shape-state space, and finding low-complexity representations has become a pressing research goal. This work tries to make a tiny step in this direction by proposing a state representation for textiles relying on the C-space of some distinctive points. A stratification of the C-space Conf_n(R^2) for n points in the cloth is derived from that of the flag manifold, and topological techniques to determine adjacencies in manipulation-centred state graphs are developed. Their algorithmic implementation permits obtaining cloth state-space representations of different granularities and tailored to particular purposes. An example of their usage to distinguish between cloth states having different manipulation affordances is provided, and ways in which the proposed state graphs can serve as a common ground to link the perception, planning and manipulation of textiles are suggested., This work is supported by the European Research Council (ERC) within the European Union Horizon 2020 Programme under grant agreement ERC-2016-ADG-741930 (CLOTHILDE: CLOTH manIpulation Learning from DEmonstrations) and by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Science-fiction: a mirror for the future of humankind

  • Torras, Carme|||0000-0002-2933-398X
Digital technologies have become part of our everyday lives and are increasingly acting as intermediaries in our workplaces and personal relationships or even substituting them. This poses a series of ethical questions that were not relevant for other types of machines and about which we have no previous experience, nor can we reliably predict how they will ultimately influence the evolution of humankind. This has led to the confluence of artificial intelligence (AI) with the humanities in an ethical debate that is starting to bear fruit, not only with the establishment of regulations and standards, but also with educational initiatives in university teaching, professional improvement, and the conformation of public opinion. Interestingly, science fiction (SF) often plays a prominent speculative role in highlighting the pros and cons of potential scenarios, thus favoring an engaging debate on AI and ethics., This work has been partly supported by the E uropean R esearch C ouncil (E R C ) under theE uropean U nion’s H orizon 2020 research and innovation programme through the projectC L O T H I L D E – C L O T H manI pulation L earning from D E monstrations (A dvanced G rantagreement N o 741930), and the S panish R esearch A gency through the M aría de M aeztu S eal ofE xcellence to I R I (M D M -2016-0656, Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Representació dels estats d’una tela basada en la derivada de la integral d’enllaç de Gauss

  • Coltraro, Franco|||0000-0002-9149-950X
  • Fontana, Josep
  • Alberich Carramiñana, Maria|||0000-0003-2749-4875
  • Borràs Sol, Júlia|||0000-0002-1662-2037
  • Torras, Carme|||0000-0002-2933-398X
La manipulació robòtica de teles és una tasca complexa a causa de l’espai d’estats de forma d’infinites dimensions dels tèxtils, que dificulta molt la percepció dels estats. En aquest article introduïm les coordenades de tela dGLI, una representació de dimensió baixa dels estats d’una tela que ens permet distingir de manera eficient una gran varietat d’estats de plegament, fet que obre les portes a mètodes d’aprenentatge eficients per a la planificació i el control de la manipulació de teles. La representació es basa en una derivada direccional de la integral d’enllaç de Gauss i ens permet representar configuracions de plegament espacials i planàries de manera consistent i unificada. Es comprova que les coordenades de tela dGLI que proposem són més encertades en la representació dels estats de la tela i significativament més sensibles als canvis en possibilitats de prensió que altres mètodes clàssics de distància de formes. Per acabar, apliquem la representació a imatges reals d’una tela i es demostra que permet identificar els diferents estats utilitzant un classificador basat en la distància., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

A topological extension of movement primitives for curvature modulation and sampling of robot motion

  • Colomé Figueras, Adrià|||0000-0001-9715-4062
  • Torras, Carme|||0000-0002-2933-398X
The version of record is available online at: https://doi.org/10.1007/s10514-021-09976-7, This paper proposes to enrich robot motion data with trajectory curvature information. To do so,we use an approximate implementation of a topological feature named writhe, which measures the curling of a closed curve around itself, and its analog feature for two closed curves, namely the linking number. Despite these features have been established for closed curves, their definition allows for a discrete calculation that is well-defined for non-closed curves and can thus provide information about how much a robot trajectory is curling around a line in space. Such lines can be predefined by a user, observed by vision or, in our case, inferred as virtual lines in space around which the robot motion is curling. We use these topological features to augment the data of a trajectory encapsulated as a Movement Primitive (MP). We propose a method to determine how many virtual segments best characterize a trajectory and then find such segments. This results in a generative model that permits modulating curvature to generate new samples, while still staying within the dataset distribution and being able to adapt to contextual variables., This work has been carried out within the project CLOTHILDE (”CLOTH manIpulation Learning from DEmonstrations”) funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Advanced Grant agreement No 741930). Research at IRI is also supported by the Spanish State Research Agency through the Mar ́ıa de Maeztu Seal of Excellence to IRI MDM-2016-0656, Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Assistive robotics: research challenges and ethics education initiatives

  • Torras, Carme|||0000-0002-2933-398X
Assistive robotics is a fast growing field aimed at helping healthcarers in hospitals, rehabilitation centers and nursery homes, as well as empowering people with reduced mobility at home, so that they can autonomously fulfill their daily living activities. The need to function in dynamic human-centered environments poses new research challenges: robotic assistants need to have friendly interfaces, be highly adaptable and customizable, very compliant and intrinsically safe to people, as well as able to handle deformable materials. Besides technical challenges, assistive robotics raises also ethical defies, which have led to the emergence of a new discipline: Roboethics. Several institutions are developing regulations and standards, and many ethics education initiatives include contents on human-robot interaction and human dignity in assistive situations. In this paper, the state of the art in assistive robotics is briefly reviewed, and educational materials from a university course on Ethics in Social Robotics and AI focusing on the assistive context are presented., Peer Reviewed
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Task-adaptive robot learning from demonstration with gaussian process models under replication

  • Arduengo García, Miguel
  • Colomé Figueras, Adrià
  • Borràs Sol, Júlia
  • Sentís Álvarez, Luis
  • Torras, Carme
© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Learning from Demonstration (LfD) is a paradigm that allows robots to learn complex manipulation tasks that can not be easily scripted, but can be demonstrated by a human teacher. One of the challenges of LfD is to enable robots to acquire skills that can be adapted to different scenarios. In this paper, we propose to achieve this by exploiting the variations in the demonstrations to retrieve an adaptive and robust policy, using Gaussian Process (GP) models. Adaptability is enhanced by incorporating task parameters into the model, which encode different specifications within the same task. With our formulation, these parameters can be either real, integer, or categorical. Furthermore, we propose a GP design that exploits the structure of replications, i.e., repeated demonstrations with identical conditions within data. Our method significantly reduces the computational cost of model fitting in complex tasks, where replications are essential to obtain a robust model. We illustrate our approach through several experiments on a handwritten letter demonstration dataset., This work has been partially funded by the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE) and by the Spanish State Research Agency through the Mar ́ıa de Maeztu Seal of Excellence to IRI [MDM-2016-0656]., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Assistive robotics: research challenges and ethics education initiatives

  • Torras, Carme
Assistive robotics is a fast growing field aimed at helping healthcarers in hospitals, rehabilitation centers and nursery homes, as well as empowering people with reduced mobility at home, so that they can autonomously fulfill their daily living activities. The need to function in dynamic human-centered environments poses new research challenges: robotic assistants need to have friendly interfaces, be highly adaptable and customizable, very compliant and intrinsically safe to people, as well as able to handle deformable materials. Besides technical challenges, assistive robotics raises also ethical defies, which have led to the emergence of a new discipline: Roboethics. Several institutions are developing regulations and standards, and many ethics education initiatives include contents on human-robot interaction and human dignity in assistive situations. In this paper, the state of the art in assistive robotics is briefly reviewed, and educational materials from a university course on Ethics in Social Robotics and AI focusing on the assistive context are presented., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Exploiting symmetries in reinforcement learning of bimanual robotic tasks

  • Amadio, Fabio
  • Colomé Figueras, Adrià
  • Torras, Carme
© 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Movement Primitives (MPs) have been widely adopted for representing and learning robotic movements using Reinforcement Learning Policy Search. Probabilistic Movement Primitives (ProMPs) are a kind of MP based on a stochastic representation over sets of trajectories, able of capturing the variability allowed while executing a movement. This approach has proved effective in learning a wide range of robotic movements, but it comes with the need of dealing with a high-dimensional space of parameters. This may be a critical problem when learning tasks with two robotic manipulators, and this work proposes an approach to reduce the dimension of the parameter space based on the exploitation of symmetry. A symmetrization method for ProMPs is presented and used to represent two movements, employing a single ProMP for the first arm and a symmetry surface that maps that ProMP to the second arm. This symmetric representation is then adopted in reinforcement learning of bimanual tasks (from user-provided demonstrations), using Relative Entropy Policy Search (REPS) algorithm. The symmetry-based approach developed has been tested in an experiment of cloth manipulation, showing a speed increment in learning the task., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Science-fiction: a mirror for the future of humankind

  • Torras, Carme
Digital technologies have become part of our everyday lives and are increasingly acting as intermediaries in our workplaces and personal relationships or even substituting them. This poses a series of ethical questions that were not relevant for other types of machines and about which we have no previous experience, nor can we reliably predict how they will ultimately influence the evolution of humankind. This has led to the confluence of artificial intelligence (AI) with the humanities in an ethical debate that is starting to bear fruit, not only with the establishment of regulations and standards, but also with educational initiatives in university teaching, professional improvement, and the conformation of public opinion. Interestingly, science fiction (SF) often plays a prominent speculative role in highlighting the pros and cons of potential scenarios, thus favoring an engaging debate on AI and ethics., This work has been partly supported by the E uropean R esearch C ouncil (E R C ) under theE uropean U nion’s H orizon 2020 research and innovation programme through the projectC L O T H I L D E – C L O T H manI pulation L earning from D E monstrations (A dvanced G rantagreement N o 741930), and the S panish R esearch A gency through the M aría de M aeztu S eal ofE xcellence to I R I (M D M -2016-0656, Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

The effects of gender and personality of robot assistants on customers’ acceptance of their service

  • Forgas Coll, Santiago
  • Huertas Garcia, Rubén
  • Andriella, Antonio
  • Alenyà Ribas, Guillem
The Covid-19 pandemic has stimulated the use of social robots in front-office services. However, some initial applications yielded disappointing results, as managers were unaware of the level of development of the robots’ artificial intelligence systems. This study proposes to adapt the Almere model to estimate the technological acceptance of service robots, which express their gender and personality, whilst assisting consumers. A 2¿×¿2 (two genders vs. two personalities) between-subjects experiment was conducted with 219 participants. Model estimation with Structural Equation Modelling confirmed seven out of eight hypotheses, and all four scenarios were estimated with Ordinary Least Squares, showing that robot gender and personality affected their technological acceptance., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

A Grasping-centered analysis for cloth manipulation

  • Borràs Sol, Júlia
  • Alenyà Ribas, Guillem
  • Torras, Carme
© 2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Compliant and soft hands have gained a lot of atten- tion in the past decade because of their ability to adapt to the shape of the objects, increasing their effectiveness for grasping. However, when it comes to grasping highly flexible objects such as textiles, we face the dual problem: it is the object that will adapt to the shape of the hand or gripper. In this context, the classic grasp analysis or grasping taxonomies are not suitable for describing textile objects grasps. This article proposes a novel definition of textile object grasps that abstracts from the robotic embodiment or hand shape and recovers concepts from the early neuroscience literature on hand prehension skills. This framework enables us to identify what grasps have been used in literature until now to perform robotic cloth manipulation, and allows for a precise definition of all the tasks that have been tackled in terms of manipulation primitives based on regrasps. In addition, we also review what grippers have been used. Our analysis shows how the vast majority of cloth manipulations have relied only on one type of grasp, and at the same time we identify several tasks that need more variety of grasp types to be executed successfully. Our framework is generic, provides a classification of cloth manipulation primitives and can inspire gripper design and benchmark construction for cloth manipulation., This workwas supported in part by the European Research Council from the EuropeanUnion Horizon 2020 Programme under Grant 741930 (CLOTHILDE: CLOTHmanIpulation Learning from DEmonstrations), in part by the Spanish StateResearch Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656), and in part by the “Ramon y Cajal” Fellowship RYC-2017-22703., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Dimensionality reduction in learning Gaussian mixture models of movement primitives for contextualized action selection and adaptation

  • Colomé Figueras, Adrià
  • Torras, Carme
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Robotic manipulation often requires adaptation to changing environments. Such changes can be represented by a certain number of contextual variables that may be observed or sensed in different manners. When learning and representing robot motion –usually with movement primitives–, it is desirable to adapt the learned behaviors to the current context. Moreover, different actions or motions can be considered in the same framework, using contextualization to decide which action applies to which situation. Such frameworks, however, may easily become large-dimensional, thus requiring to reduce the dimensionality of the parameters space, as well as the amount of data needed to generate and improve the model over experience. In this paper, we propose an approach to obtain a generative model from a set of actions that share a common feature. Such feature, namely a contextual variable, is plugged into the model to generate motion. We encode the data with a Gaussian Mixture Model in the parameter space of Probabilistic Movement Primitives (ProMPs), after performing Dimensionality Reduction (DR) on such parameter space. We append the contextual variable to the parameter space and obtain the number of Gaussian components, i.e., different actions in a dataset, through Persistent Homology. Then, using multimodal Gaussian Mixture Regression (GMR), we can retrieve the most likely actions given a contextual situation and execute them. After actions are executed, we use Reward-Weighted Responsibility GMM (RWR-GMM) update the model after each execution. Experimentation in 3 scenarios shows that the method drastically reduces the dimensionality of the parameter space, thus implementing both action selection and adaptation to a changing situation in an efficient way., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

A topological extension of movement primitives for curvature modulation and sampling of robot motion

  • Colomé Figueras, Adrià
  • Torras, Carme
The version of record is available online at: https://doi.org/10.1007/s10514-021-09976-7, This paper proposes to enrich robot motion data with trajectory curvature information. To do so,we use an approximate implementation of a topological feature named writhe, which measures the curling of a closed curve around itself, and its analog feature for two closed curves, namely the linking number. Despite these features have been established for closed curves, their definition allows for a discrete calculation that is well-defined for non-closed curves and can thus provide information about how much a robot trajectory is curling around a line in space. Such lines can be predefined by a user, observed by vision or, in our case, inferred as virtual lines in space around which the robot motion is curling. We use these topological features to augment the data of a trajectory encapsulated as a Movement Primitive (MP). We propose a method to determine how many virtual segments best characterize a trajectory and then find such segments. This results in a generative model that permits modulating curvature to generate new samples, while still staying within the dataset distribution and being able to adapt to contextual variables., This work has been carried out within the project CLOTHILDE (”CLOTH manIpulation Learning from DEmonstrations”) funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Advanced Grant agreement No 741930). Research at IRI is also supported by the Spanish State Research Agency through the Mar ́ıa de Maeztu Seal of Excellence to IRI MDM-2016-0656, Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Ethics of social robotics: individual and societal concerns and opportunities

  • Torras, Carme
© 2024 by the author(s). This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. See credit lines of images or other third-party material in this article for license information., Focus on the ethics of a given technology tends to lag far behind its development. This lag has been particularly acute in the case of artificial intelligence, whose accelerated deployment in a wide range of domains has triggered unprecedented attention on the risks and consequences for society at large, leading to a myriad of ethics regulations, which are difficult to coordinate and integrate due to their late appearance. The very nature of social robots forces their deployment to occur at a much slower pace, providing an opportunity for a profound reflection on ethics, which is already happening in multidisciplinary teams. This article provides a personal view of the ethics landscape, centered on the particularities of social robotics, with the main issues being ordered along two axes (individual and societal) and grouped into eight categories (human dignity, human autonomy, robot transparency, emotional bonding, privacy and safety, justice, freedom, and responsibility). This structure stems from the experience of developing and teaching a university course on ethics in social robotics, whose pedagogical materials are freely available., This work was partly supported by the European Research Council through the project CLOTHILDE (Cloth Manipulation Learning from Demonstrations; advanced grant agreement 741930) and by the European Union's Horizon Europe program through the project SoftEnable (HORIZON-CL4-2021-DIGITAL-EMERGING-01-101070600)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Benchmarking bimanual cloth manipulation

  • Garcia Camacho, Irene
  • Lippi, Martina
  • Welle, Michael C.
  • Yin, Hang
  • Antonova, Rika
  • Varava, Anastasiia
  • Borràs Sol, Júlia
  • Torras, Carme
  • Marino, Alessandro
  • Alenyà Ribas, Guillem
  • Kragic, Danica
© 2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Cloth manipulation is a challenging task that, despite its importance, has received relatively little attention compared to rigid object manipulation. In this paper, we provide three benchmarks for evaluation and comparison of different approaches towards three basic tasks in cloth manipulation: spreading a tablecloth over a table, folding a towel, and dressing. The tasks can be executed on any bimanual robotic platform and the objects involved in the tasks are standardized and easy to acquire. We provide several complexity levels for each task, and describe the quality measures to evaluate task execution. Furthermore, we provide baseline solutions for all the tasks and evaluate them according to the proposed metrics., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Social networks and robot companions: technology, ethics and science fiction

  • Torras, Carme
Information technologies have become part of our everyday lives and are increasingly acting as intermediaries in our workplaces and personal relationships or even substituting them. This growing interaction with machines poses several questions about which we have no previous experience, nor can we reliably predict how they will influence the evolution of society. This has led to the convergence of technoscience and humanities in an ethical debate that is starting to bear fruit, not only with the setting of regulations and standards, but also with educational initiatives in university teaching, professional improvement, and the conformation of public opinion. Interestingly, science fiction often plays a prominent speculative role in highlighting the pros and cons of potential scenarios., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Benchmarking the sim-to-real gap in cloth manipulation

  • Blanco Mulero, David
  • Barbany Mayor, Oriol
  • Alcan, Gokhan
  • Colomé Figueras, Adrià
  • Torras, Carme
  • Kyrki, Ville
Realistic physics engines play a crucial role for learning to manipulate deformable objects such as garments in simulation. By doing so, researchers can circumvent challenges such as sensing the deformation of the object in the real-world. In spite of the extensive use of simulations for this task, few works have evaluated the reality gap between deformable object simulators and real-world data. We present a benchmark dataset to evaluate the sim-to-real gap in cloth manipulation. The dataset is collected by performing a dynamic as well as a quasi-static cloth manipulation task involving contact with a rigid table. We use the dataset to evaluate the reality gap, computational time, and simulation stability of four popular deformable object simulators: MuJoCo, Bullet, Flex, and SOFA. Additionally, we discuss the benefits and drawbacks of each simulator. The benchmark dataset is open-source. Supplementary material, videos, and code, can be found at https://sites.google.com/view/cloth-sim2real-benchmark., This work was supported in part by Academy of Finland under Grant 317020, in part by Business Finland (decision 9249/31/2021), in part by EU’s H2020 Programme through the CLOTHILDE Project under Grant ERC-2016-ADG-741930, in part by EU’s Horizon Europe Programme through the SoftEnable Project under Grant Agreement 101070600, and in part by MCIN/AEI/10.13039/501100011033 through CHLOE-GRAPH Project under Grant PID2020-118649RB-I00., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Dimensionality reduction for dynamic movement primitives and application to bimanual manipulation of clothes

  • Colomé Figueras, Adrià
  • Torras, Carme
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Dynamic Movement Primitives (DMPs) are nowadays widely used as movement parametrization for learning robot trajectories, because of their linearity in the parameters, rescaling robustness and continuity. However, when learning a movement with DMPs, a very large number of Gaussian approximations needs to be performed. Adding them up for all joints yields too many parameters to be explored when using Reinforcement Learning (RL), thus requiring a prohibitive number of experiments/simulations to converge to a solution with a (locally or globally) optimal reward. In this paper we address the process of simultaneously learning a DMP-characterized robot motion and its underlying joint couplings through linear Dimensionality Reduction (DR), which will provide valuable qualitative information leading to a reduced and intuitive algebraic description of such motion. The results in the experimental section show that not only can we effectively perform DR on DMPs while learning, but we can also obtain better learning curves, as well as additional information about each motion: linear mappings relating joint values and some latent variables., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Deformable surface reconstruction via Riemannian metric preservation

  • Barbany Mayor, Oriol
  • Colomé Figueras, Adrià
  • Torras, Carme
Estimating the pose of an object from a monocular image is a fundamental inverse problem in computer vision. Due to its ill-posed nature, solving this problem requires incorporating deformation priors. In practice, many materials do not perceptibly shrink or extend when manipulated, constituting a reliable and well-known prior. Mathematically, this translates to the preservation of the Riemannian metric. Neural networks offer the perfect playground to solve the surface reconstruction problem as they can approximate surfaces with arbitrary precision and allow the computation of differential geometry quantities. This paper presents an approach for inferring continuous deformable surfaces from a sequence of images, which is benchmarked against several techniques and achieves state-of-the-art performance without the need for offline training. Being a method that performs per-frame optimization, our method can refine its estimates, contrary to those based on performing a single inference step. Despite enforcing differential geometry constraints at each update, our approach is the fastest of all the tested optimization-based methods., This work is part of the project CLOTHILDE (‘‘CLOTH manIpulation Learning from DEmonstrations’’) which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (Advanced Grant agreement No. 741930)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Are preferences useful for better assistance? A physically assistive robotics user study

  • Canal Camprodon, Gerard
  • Torras, Carme
  • Alenyà Ribas, Guillem
© 2021 Copyright held by the owner/author(s)., Assistive Robots have an inherent need of adapting to the user they are assisting. This is crucial for the correct development of the task, user safety, and comfort. However, adaptation can be performed in several manners. We believe user preferences are key to this adaptation. In this paper, we evaluate the use of preferences for Physically Assistive Robotics tasks in a Human-Robot Interaction user evaluation. Three assistive tasks have been implemented consisting of assisted feeding, shoe-fitting, and jacket dressing, where the robot performs each task in a different manner based on user preferences. We assess the ability of the users to determine which execution of the task used their chosen preferences (if any). The obtained results show that most of the users were able to successfully guess the cases where their preferences were used even when they had not seen the task before. We also observe that their satisfaction with the task increases when the chosen preferences are employed. Finally, we also analyze the user’s opinions regarding assistive tasks and preferences, showing promising expectations as to the benefits of adapting the robot behavior to the user through preferences., This work has been supported by the ERC project Clothilde (ERC-2016-ADG-741930), the HuMoUR project (Spanish Ministry of Science and Innovation TIN2017-90086-R) and by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656). Gerard Canal has also been supported by the Spanish Ministry of Education, Culture and Sport by the FPU15/00504 doctoral grant and the CHIST-ERA project COHERENT (EPSRC EP/V062506/1)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Gendered human–robot Interactions in services

  • Forgas Coll, Santiago
  • Huertas Garcia, Rubén
  • Andriella, Antonio
  • Alenyà Ribas, Guillem
© The Author(s) 2023, The outbreak of Covid-19 precipitated the use of service robots in customer-facing services as a replacement for employ- ees to avoid human-to-human contact. However, this development has not resolved the debate as to whether robots should be characterized with gender attributes or simply be genderless. This study explores whether endowing a robot with gender attributes makes it more acceptable as a service provider among stated men and women. To this end, an experiment was conducted at a public fair in which a gendered robot simulated the provision of a service to customers, which consisted of offering them advice, hints, and messages of encouragement to help complete a eudaemonic puzzle. A parsimonious version of the Almere model was used to estimate acceptance of the technology. The findings reveal that for both stated men and women, the main drivers for accepting the female-coded robot are perceived usefulness and social influence, although women attach greater importance to social influence. For the male-coded robot, perceived usefulness and social influence are the main arguments for women, while for men they are enjoyment, perceived usefulness and, negatively, ease of use. In addition, different indirect effects between stated sexes are also identified. In summary, men and women consider different factors when accepting robots of each gender., This work was partially funded by the European Union’s Horizon 2020 programme under ERC Advanced Grant CLOTHILDE (no. 741930) and under the Marie Skłodowska-Curie grant agreement (no. 712949) (TECNIOspring PLUS); by MCIN/ AEI /https://doi.org/10.13039/501100011033 and by the “European Union NextGenerationEU/PRTR” under the project ROB-IN (PLEC2021-007859); by the Research Council of Norway under the project SECUROPS (INT-NO/0875); and by the “European Union NextGenerationEU/PRTR” through CSIC’s Thematic Platforms (PTI + Neuro-Aging)., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Model predictive control for dynamic cloth manipulation: parameter learning and experimental validation

  • Luque Acera, Adrià
  • Parent Alonso, David
  • Colomé Figueras, Adrià
  • Ocampo-Martínez, Carlos
  • Torras, Carme
© 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Robotic cloth manipulation is a challenging problem for robotic systems. Textile items can adopt multiple configurations and shapes during their manipulation. Hence, robots should not only understand the current configuration of the item but also be able to predict its future possible behaviours and perform real-time control during manipulation. This paper addresses the problem of indirectly controlling the configuration of certain points of a textile object, by applying actions on other parts of it through the use of a Model Predictive Control (MPC) strategy. MPC allows to foresee the behavior of indirectly controlled points, while satisfying physical/operational constraints. This is done by first identifying the optimal control signals that may constitute the desired future cloth configuration. After that, a dynamic model of the item will be used and sensor data will allow to update the belief on the object’s state and close the loop. This paper investigates how grasping the upper corners of a square piece of cloth can allow to track a reference trajectory of the pieces’ lower corners. To do so, we propose and validate a linear cloth model that allows solving the MPC optimization problem in real time. Reinforcement Learning (RL) techniques are used to learn the parameters of the proposed cloth model and to tune the resulting MPC. The full control scheme was implemented and executed in a real robot, obtaining accurate tracking results in adverse conditions., This work was developed in the context of the project CLOTHILDE (”CLOTH manIpulation Learning from DEmonstrations”), which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programmee (Advanced Grant agreement No 741930). The work is also supported by Project PID2020- 118649RB-I00 funded by MCIN/ AEI /10.13039/501100011033., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Representació dels estats d’una tela basada en la derivada de la integral d’enllaç de Gauss

  • Coltraro, Franco
  • Fontana, Josep
  • Alberich Carramiñana, Maria
  • Borràs Sol, Júlia
  • Torras, Carme
La manipulació robòtica de teles és una tasca complexa a causa de l’espai d’estats de forma d’infinites dimensions dels tèxtils, que dificulta molt la percepció dels estats. En aquest article introduïm les coordenades de tela dGLI, una representació de dimensió baixa dels estats d’una tela que ens permet distingir de manera eficient una gran varietat d’estats de plegament, fet que obre les portes a mètodes d’aprenentatge eficients per a la planificació i el control de la manipulació de teles. La representació es basa en una derivada direccional de la integral d’enllaç de Gauss i ens permet representar configuracions de plegament espacials i planàries de manera consistent i unificada. Es comprova que les coordenades de tela dGLI que proposem són més encertades en la representació dels estats de la tela i significativament més sensibles als canvis en possibilitats de prensió que altres mètodes clàssics de distància de formes. Per acabar, apliquem la representació a imatges reals d’una tela i es demostra que permet identificar els diferents estats utilitzant un classificador basat en la distància., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Controlled Gaussian process dynamical models with application to robotic cloth manipulation

  • Amadio, Fabio
  • Delgado Guerrero, Juan Antonio
  • Colomé Figueras, Adrià
  • Torras, Carme
Over the last years, significant advances have been made in robotic manipulation, but still, the handling of non-rigid objects, such as cloth garments, is an open problem. Physical interaction with non-rigid objects is uncertain and complex to model. Thus, extracting useful information from sample data can considerably improve modeling performance. However, the training of such models is a challenging task due to the high-dimensionality of the state representation. In this paper, we propose Controlled Gaussian Process Dynamical Models (CGPDMs) for learning high-dimensional, nonlinear dynamics by embedding them in a low-dimensional manifold. A CGPDM is constituted by a low-dimensional latent space, with an associated dynamics where external control variables can act and a mapping to the observation space. The parameters of both maps are marginalized out by considering Gaussian Process priors. Hence, a CGPDM projects a high-dimensional state space into a smaller dimension latent space, in which it is feasible to learn the system dynamics from training data. The modeling capacity of CGPDM has been tested in both a simulated and a real scenario, where it proved to be capable of generalizing over a wide range of movements and confidently predicting the cloth motions obtained by previously unseen sequences of control actions., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work was partially developed in the context of the project CLOTHILDE (“CLOTH manIpulation Learning from DEmonstrations”), which has received funding from ERC under the European Union’s Horizon 2020 research and innovation program (Advanced Grant agreement No 741930). This work has also received funding from project CHLOE-GRAPH (PID2020-118649RB-I00) funded by MCIN/ AEI /10.13039/501100011033., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Social robot-delivered customer-facing services: an assessment of the experience

  • Forgas Coll, Santiago
  • Huertas Garcia, Rubén
  • Andriella, Antonio
  • Alenyà Ribas, Guillem
The ability to install social intelligence protocols in robots in order for them to exhibit conversational skills has made them ideal tools for delivering services with a high cognitive and low emotional load. Little is known about how this capability influences the customer experience and the intention to continue receiving these services. Experiences were assessed in a study simulating customer-facing service delivery, and the constructs of the technology readiness index and stated gender were analysed as possible moderators in a quasi-experiment. Hedonic quality was the most relevant factor explaining attitude, and attitude explained intention to use as well as social influence. As for the constructs of technological readiness and gender, optimism and innovativeness seem to be the most likely candidates for moderating the other variables. The most optimistic and the most innovative route would be for the main actors to continue adapting to social robot technology in the future., This work was partially funded by the European Union’s Horizon 2020 under ERC Advanced Grant CLOTHILDE (no. 741930); by MCIN/ AEI /10.13039/501100011033 and by the ”European Union NextGenerationEU/PRTR” under the project ROB-IN (PLEC2021-007859); by the Research Council of Norway under the project SECUROPS (INT-NO/0875); and by the ”European Union NextGenerationEU/PRTR” through CSIC’s Thematic Platforms (PTI+ Neuro-Aging)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

The effects of gender and personality of robot assistants on customers' acceptance of their service

  • Forgas Coll, Santiago
  • Huertas García, Rubén
  • Andriella, Antonio
  • Alenyà, Guillem
The Covid-19 pandemic has stimulated the use of social robots in front-office services. However, some initial applications yielded disappointing results, as managers were unaware of the level of development of the robots' artificial intelligence systems. This study proposes to adapt the Almere model to estimate the technological acceptance of service robots, which express their gender and personality, whilst assisting consumers. A 2 × 2 (two genders vs. two personalities) between-subjects experiment was conducted with 219 participants. Model estimation with Structural Equation Modelling confirmed seven out of eight hypotheses, and all four scenarios were estimated with Ordinary Least Squares, showing that robot gender and personality affected their technological acceptance.
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Topological representation of cloth state for robot manipulation: deriving the configuration space of a rectangular cloth

  • Strazzeri, Fabio
  • Torras, Carme
The version of record is available online at: https://doi.org/10.1007/s10514-021-09968-7, Forty years ago the notion of configuration space (C-space) revolutionised robot motion planning for rigid and articulated objects. Despite great progress, handling deformable materials has remained elusive because of their infinite-dimensional shape-state space, and finding low-complexity representations has become a pressing research goal. This work tries to make a tiny step in this direction by proposing a state representation for textiles relying on the C-space of some distinctive points. A stratification of the C-space Conf_n(R^2) for n points in the cloth is derived from that of the flag manifold, and topological techniques to determine adjacencies in manipulation-centred state graphs are developed. Their algorithmic implementation permits obtaining cloth state-space representations of different granularities and tailored to particular purposes. An example of their usage to distinguish between cloth states having different manipulation affordances is provided, and ways in which the proposed state graphs can serve as a common ground to link the perception, planning and manipulation of textiles are suggested., This work is supported by the European Research Council (ERC) within the European Union Horizon 2020 Programme under grant agreement ERC-2016-ADG-741930 (CLOTHILDE: CLOTH manIpulation Learning from DEmonstrations) and by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Learning in autonomous and intelligent systems: Overview and biases from data sources

  • Jiménez Schlegl, Pablo
© 2021 CSIC. Este es un artículo de acceso abierto distribuido bajo los términos de la licencia de uso y distribu-ción Creative Commons Reconocimiento 4.0 Internacional (CC BY 4.0)., Autonomous and Intelligent Systems (A/IS, to adhere to the terminology of the IEEE Ethically Aligned Design report) can gather their knowledge by different means and from different sources. In principle, learning algorithms are neutral; rather, it is the data they are fed during the learning period that can introduce biases or a specific ethical orientation. Human control over the learning process is more straightforward in learning from demonstration, where data sources are restricted to the choices of the demonstrator (or teacher), but even in unsupervised versions of reinforcement learning, biases are present via the definition of the reward function. In this paper we provide an overview of learning paradigms of artificial systems: supervised and unsupervised methods, with the most striking examples in each category, without too much technical detail. Furthermore, we describe the types of data sources that are presently available and in use by the robotics community. We also focus on observable bias in image datasets and originated by human annotation. We point at quite recent research on bias in social robot navigation and end with a brief reflection about ambient influences on future learning robots., This work has been partially funded by the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE) and by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI[MDM-2016-0656] and the project HuMoUR TIN2017-90086-R., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

A Versatile Gripper for Cloth Manipulation

  • Donaire Cónsul, Sònia
  • Borràs Sol, Júlia
  • Alenyà Ribas, Guillem
  • Torras, Carme
Cloth manipulation has been mostly advancing in perception and modeling methods for cloth state estimation and grasping point detection. In comparison, less attention has been put on end-effector design. Indeed, most implementations use a parallel gripper that can only perform pinch grasps. Instead, a more versatile set of possible grasp types could ease many tasks by providing more support to certain parts of the object, as well as make feasible tasks that become very complex when executed with only pinch grasps. We present a versatile gripper design which, besides the common open-close thumb+finger(s) feature, it has a couple of reconfiguration degrees of freedom that offer, first, a wide base plane to provide support and, second, variable friction surface on the thumb tip. Our gripper can execute a versatile set of grasps that ease some complex tasks such as pick and place folded clothes or fold in the air. In addition, the variable friction mechanism enables a more robust pinch-and-slide manipulation to trace cloth edges. Our evaluation shows the gripper potential to execute a wide variety of cloth manipulation tasks., This work receives funding from the European Union Horizon 2020 Programme under grant agreement no. 741930 (CLOTHILDE), Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656) and the “Ramon y Cajal” Fellowship RYC-2017-22703, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Generating predicate suggestions based on the space of plans: an example of planning with preferences

  • Canal Camprodon, Gerard
  • Torras, Carme
  • Alenyà Ribas, Guillem
The version of record is available online at: https://dx.doi.org/10.1007/s11257-022-09327-w, Task planning in human–robot environments tends to be particularly complex as it involves additional uncertainty introduced by the human user. Several plans, entailing few or various differences, can be obtained to solve the same given task. To choose among them, the usual least-cost plan criteria is not necessarily the best option, because here, human constraints and preferences come into play. Knowing these user preferences is very valuable to select an appropriate plan, but the preference values are usually hard to obtain. In this context, we propose the Space-of-Plans-based Suggestions (SoPS) algorithms that can provide suggestions for some planning predicates, which are used to define the state of the environment in a task planning problem where actions modify the predicates. We denote these predicates as suggestible predicates, of which user preferences are a particular case. The first algorithm is able to analyze the potential effect of the unknown predicates and provide suggestions to values for these unknown predicates that may produce better plans. The second algorithm is able to suggest changes to already known values that potentially improve the obtained reward. The proposed approach utilizes a Space of Plans Tree structure to represent a subset of the space of plans. The tree is traversed to find the predicates and the values that would most increase the reward, and output them as a suggestion to the user. Our evaluation in three preference-based assistive robotics domains shows how the proposed algorithms can improve task performance by suggesting the most effective predicate values first., This work has been partially supported by the ERC project Clothilde (ERC-2016-ADG-741930); by MCIN/AEI/10.13039/501100011033 under the project CHLOE-GRAPH (PID2020-118649RB-l00); by the European Union NextGenerationEU/PRTR under the project ROB-IN (PLEC2021-007859); and by the CHIST-ERA project COHERENT (EPSRC EP/V062506/1), Peer Reviewed, Postprint (published version)