Semánticas Argumentativas Basadas en Asignación de Estados

Social robots are robots which mainly interact with people using speech that takes social context into consideration as well as facial expressions, physical gestures, and other social cues. They can be used to help people by acting as teachers, rehabilita- tive coaches, assistants for the physically impaired, carers for the elderly, and other non-contact assistive roles. In addition, they often function as research platforms for investigating features about specific modes of social interaction [Dautenhahn, 2007] [Fong et al., 2003a].

Evaluating and classifying social robotics

Feil-Seifer and Matari´c described a taxonomy of socially assistive robotics, a subset of assistive robots which aid their human users via their social interactions. They categorized the robots according to such factors as their user populations, examples of the tasks the robots perform, the sophistication of the interactions, and the role of the robot in relation to other humans with whom it will work. In addition, they also described ways that such robots should be evaluated [Feil-Seifer and Matari´c, 2005]. Feil-Seifer, Skinner, and Matari´c later elaborated on different benchmarks that could be used to evaluate socially assistive robotics, such as the safety and scalability of the robotic technology itself; the suitably-trusted degree of autonomy, the best usage of the robot’s imitative capabilities, and the impact of data privacy on the success of the robot’s social interactions; how well the robot had achieved its desired social identity and whether the robot’s social understanding of people had helped in its performance of necessary tasks; and how the assistive technology itself had impacted the quality of care given to the users, the users’ overall quality of life, and, in the cases where the robot will be supervised by human caregivers, the ease of the human caregivers’ jobs [Feil-Seifer et al., 2007]. In another review of socially assistive robotics, Tapus, Matari´c, and Scassellati described the biggest challenges in socially assistive robotics: the role of embodiment in people’s interactions with robots; how a robot’s interactions should be matched to suit the user’s personality; how the robots’ simulations of empathy could affect their interactions with people; how a robot could monitor the level of engagement that a human had in their interactions; how a robot could learn the behaviours of its users and adapt its approaches over time to maintain social engagement; and how easily the skills learned with the robot could transfer into other contexts

Paro

Shibata first described Paro, the robot that featured very prominently in his re- search, as a robotic pet or artificial emotional creature. The robot’s emotions would gradually change over time according to the data it got from its visual, audio, and tactile sensors, and the robot would express these emotions in its actions [Shibata et al., 1996]. When Paro was used in robot-assisted therapy at a children’s hospital, participants reported that they felt happier while playing with the interactive form of the robot than during sessions when the robot was a “stuffed animal” and did not respond to any stimuli. The robot also seemed to act as a social mediator and conversation piece for the children at the hospital, and anecdotal evidence seemed to suggest that children with autism also responded to it [Shibata et al., 2001]. When Paro was later used in free-form group robot interaction sessions at a nursing home, the participants’ questionnaires showed that they felt more energetic after interact- ing with the robot than they did beforehand. Additionally, some participants also spoke to Paro as well as touched the robot more after the final interaction session than they did after the first one [Wada et al., 2002]. Physiological data in the form of urinary analyses also showed that the patients were better able to deal with stress during the weeks that they interacted with Paro, and the staff at the nursing home also reported lower feelings of burnout among themselves during the weeks that the patients interacted with Paro [Wada et al., 2004]. When a similar longitudinal study was conducted, the participants’ questionnaires consistently showed that they were happier after interacting with Paro for a 5-month period. Caregivers also noted that the participants looked forward to interacting with the robot, interacted more with their peers during the robot therapy sessions, and were happier as well as more energetic while interacting with Paro [Wada et al., 2005].

Probo

Saldien and Goris’s robot Probo was designed with the appearance of a child-friendly furry mastodon or mammoth, and was meant to be used as a robot companion and communication aid for children in hospitals. In addition to the robot being able to keep children entertained with its interactive behaviours, the screen on the robot’s stomach is also meant to allow children to communicate with the outside world using standard video-conferencing techniques as well as explain medical procedures to children using multimedia techniques in order to alleviate their fears [Saldien et al., 2008b]. Because the Probo has a very expressive face with eyes, eyelids, eyebrows, a mouth, and a trunk-like nose, the robot is able to express a fair number of basic emotions. Probo’s emotions are represented as vectors in a two dimensional space bound by a unit circle, with one axis for valence (happy/sad), one axis for arousal (surprised/tired), and the length of the vector as the intensity of the emotion. When participants were asked to describe Probo’s various facial expressions, each of which represented a unique emotional state of the robot, by looking at pictures of a virtual representation of Probo’s face, children successfully recognized five of the eight displayed emotions while adults successfully recognized six of the eight emotions [Saldien et al., 2008a] [Goris et al., 2008]. Furthermore, because Probo is meant to give soft hugs to children in a safe manner, the robot uses soft or springy materials whenever possible and also uses a novel, extensively-tested set of compliant acutators which use Bowden cables and non-backdriveable servos to either transmit rotational motion or a pulling force across a large distance [Goris et al., 2011].

IROMEC

The European project IROMEC (Interactive RObotic MEdiators as Companions) is aimed at designing a robotic play companion for children with limited play skills due to physical or mental impairments, since play has been recognized as an activity that helps children achieve their full learning potential and promotes social interaction.

The IROMEC robot was designed so that children could play with it in a variety of different play scenarios that are meant to have specific therapeutic or beneficial effects [Kronreif and Prazak-Aram, 2008]. In addition to the project using the ICF-CY (International Classification of Functioning and Disability, Children and Youth), a compendium developed by the World Health Organization to describe and code all health-related experiences, as a set of guidelines for designing the robot, evaluating the robot’s interactions with children, and outlining the robot’s play scenarios [Besio et al., 2008], the project has also used interviews and focus groups involving therapists, teachers, care-givers, and parents from each group of children that the project is meant to help. By gathering information about the cultural, social, emotional, and functional implications of different design choices for the robot, the IROMEC project has been able to design a modular robotic system that should address the needs of many different groups of children with disabilities [Marti et al., 2009]. Specifically, to address the needs of children with autism for help with social interaction and social communication, the IROMEC robot can play simple games involving imitation, sensory rewards, and turn-taking. Furthermore, because children with autism tend to be a diverse group of individuals with different manifestations of the same impairment and very particular preferences, the robot’s games can be tweaked and modified to focus on the specific issues of each child and its hardware is designed to be plug-and-play to allow the children to participate in the games using various devices and input methods. However the robot and its games are tailored, it will still help children with autism and other children with disabilities in their social, cognitive, sensory, motor, communication, and emotional development [Ferrari et al., 2009].