ANÁLISIS DE ESTRUCTURA DE LA CABINA DE CROMADO

Socially Assistive Robots (SARs) provide the potential to be applied as a social output component for traditional ITSs. They can be seen as a combination of two related fields intersecting each other (see Figure3.2). First, Assistive Robots (ARs) that can, for instance, help humans by lifting objects in their private homes or factory environments, or support them in physical rehabilitation, e.g., in form of a wheelchair robot (Simpson and Levine,1997, cf.Feil-Seifer and Mataric,2005). Hence, their focus is more on physically assisting humans. In contrast, the goal of so-called Social Robots or Socially Interactive Robots (SIRs) is to establish and maintain a good social interaction with people. They can

3.2. Socially Assistive Robots

ARs SIRs

SARs

Figure 3.2: The field of Socially Assistive Robots (SARs) can be seen the intersection of Socially Interac- tive Robots (SIRs) and Assistive Robots (ARs).

take several roles and can act as peers or partners with various shapes (cf.Fong et al.,2003;Feil-Seifer and Mataric,2005, see also Section3.2.1).

The field of SARs combines parts of both, establishing and maintaining a social interaction, while assisting people in various life or job settings. The latter is also the key difference to SIR. While the goal of SIR is the social interaction itself, SARs aim for establishing a social interaction to provide assis- tance and to achieve measurable progress. Furthermore, instead of assisting the user physically, which is the goal of ARs, SARs focus more on assisting through their social presence and socially supportive behaviors (cf.Feil-Seifer and Mataric,2005). For example, SARs can provide social and personalized interaction, e.g., by offering motivational and engaging long-term support for the user, to assist in var- ious tasks, such as elderly care, rehabilitation or learning (cf.Feil-Seifer and Mataric,2005;Tapus et al.,

2007;Fasola and Matarić,2013;Matarić,2014;Clabaugh et al.,2015;Gordon and Breazeal,2015). How- ever, this combination also results in the necessity to address the challenges provided by both fields. On the one hand, SARs have to face the situation-dependent challenges of each task in which they should provide assistance. On the other hand, they have to handle the new challenges arising from interacting with social individuals, such as human beings.

3.2.1 Challenges in Social Robotics

The terms of Social Robotics, Social Robots or Socially Interactive Robot have become synonyms for an upcoming field of research in the past decades. Robots are not only meant to do assembly work in factories anymore but also start to arrive in our daily lives. But this new field of application also rises new challenges, initiated through the necessity to interact with social beings. For example, they have to recognize and understand the user while acting in line with social rules and norms. This is also reflected in the definitions of Social Robots, which focus not only on the robot’s task and role but also on the challenges and requirements they have to fulfill. For instance, Social Robots can be defined as follows:

“A social robot is an autonomous or semi-autonomous robot that interacts and com- municates with humans by following the behavioral norms expected by the people with whom the robot is intended to interact.”

This definition already highlights the new abilities a social robot has to have. First, it needs to be able to interact and communicate with humans, preferably in a human like and natural way or at least in a fashion that is understandable by humans. Thus, it needs modalities to produce verbal output and/or non-verbal behaviors, such as nodding or pointing. Second, while doing so, it has to choose its actions based on behavioral norms and rules accepted by humans, which are quite more complex compared to the set of rules applied in a controlled factory environment. Finally, it has to behave semi- autonomously or even completely autonomously.

Del Moral et al.(2009) extended this definition by also adding the environment and physical con- straints of the robot:

“A Social Robot is an autonomous motion device equipped with sensors, actuators and interfaces (robot) that interacts and communicates with humans following some expected behavior rules, which are founded on the robot physical properties and the environment within it is embedded, mainly taking into account the needs of the people with witch it is meant to interact”

(del Moral et al.,2009, p. 5) Thus, if a social robot is also mobile, e.g., by moving on wheels or even on its own legs, it needs further sensors to be aware of its environment and also extended navigation skills to be able to fulfill its tasks without interfering or harming its interaction partners. Additionally, it needs appropriate actuators and interfaces to be able to interact with its environment and also with humans.

Although the navigation through social environments and the interaction with humans yield a lot of new challenges, the application of social robots also provides a lot of benefits. In contrast to an interaction with a tablet screen or a PC, the interaction with a social robot feels more natural to humans. A robot, for instance, can use common non-verbal cues, such as eye-gaze to establish joint attention, nod to show agreement or other types of gestures. However, one might argue that these behaviors can also be used by virtual agents, but research has shown that robots are perceived more helpful, credible, informative and enjoyable to interact with (Kidd and Breazeal,2004;Wainer et al.,2007).

3.2.2 Fields of Application

Although a social interaction increases the difficulty of developing social robots and, thus, also SARs, they were already applied in a variety of settings. For example, in health or convalescent care they can provide social support by distracting or engaging patients during their recovering periods in hospi- tals (Saldien et al.,2006) or support and teach them to handle their diseases (Henkemans et al.,2013;

Broadbent et al.,2018). The ALIZ-E project1, for instance, used the Nao robot (see Figure3.3a) to provide personalized health education for children with diabetes (Henkemans et al.,2013). Although non-social robots have also been shown to yield good results in rehabilitation (Lo et al.,2010) or health care (Davies,2016), there might be still a big potential to improve their effectiveness with a social com- ponent by making the therapeutic process more enjoyable (cf.Matarić et al.,2007).

3.2. Socially Assistive Robots

(a) Nao2

Figure 3.3: Socially Assistive Robots used in health (a,b) and elderly care (c,d,e).

Another field of application is the elderly care in which SARs can be used to entertain people or to help them with their mental health issues, while trying to maintain their independence as long as possi- ble (cf.Libin and Cohen-Mansfield,2004;Feil-Seifer and Mataric,2005;Banks et al.,2008;Lehmann et al.,2013;Broadbent et al.,2014;Jenkins and Draper,2014;Chang and Šabanović,2015;Orejana et al.,

2015). However, the elderly are often distrustful and, thus, hesitate or even decline to interact with a robot (cf.Miehle et al.,2019). In these cases the robot’s social abilities can be used to lower their hesita- tion and to raise their acceptance of the robot’s role and assistance, which, in fact, was already verified in several studies and even in long-term interactions (e.g.,Broadbent et al.,2014;Orejana et al.,2015). But this requires to design the robots and their behaviors carefully to allow for a significant increase in the quality of life or medication adherence in this particular setting (cf.Broadbent et al.,2014). How- ever, despite this impediment several robots with various shapes were already applied to this setting. They are ranging from large machine like robots, such as NurseBot Pearl (see Figure??) that is used to relieve nurses in their day to day activities by reminding the elderly to take their medication and guiding them around the environment (Montemerlo et al.,2002;Pineau et al.,2003), down to small pet-like robots, such as NeCoRo (Libin and Cohen-Mansfield,2004, see Figure??), AIBO (Banks et al.,

2008, see Figure??) or the seal shaped therapeutic robot called PARO (Chang and Šabanović,2015, see Figure??). Especially, these cute robots seem to have a great positive influence on the interaction with the elderly by encouraging social interactions, reducing their stress and loneliness and improving their mood (Libin and Cohen-Mansfield,2004;Šabanović et al.,2013;Aminuddin et al.,2016).

Moreover, SARs are also used as therapeutic tools to support people suffering from cognitive or social disorders, in particular, Autism Spectrum Disorder (ASD) (cf.Dautenhahn and Werry,2004). Here, they are applied to support the development of life skills to allow for higher independence and to reduce behaviors, which might interfere with this goal (cf.Begum et al.,2016). One of the most prominent aspects patients with ASD are often struggling with is to establish and maintain a social interaction with other people and, hence, this is one of the major topics addressed with SARs. Van- derborght et al.(2012), for instance, used the robot Probo (see Figure??) to increase the social skills of

2_{https://www.softbankrobotics.com/emea/en/robots/nao}

3_{https://www.cmu.edu/cmtoday/issues/dec-2004-issue/feature-stories/human-health/index.html} 4_{http://parorobots.com}

5_{http://www.megadroid.com/Robots/necoro.htm} 6_{https://us.aibo.com/}

Figure 3.4: Socially Assistive Robots used to support people with cognitive or social disorders.

children suffering from ASD within a story telling setting and, indeed, their results show that Probo is able to improve their social performance in specific situations (Vanderborght et al.,2012). Another robot that is applied in this scope is called Keepon (see Figure??). Although it is very simplistic and only capable of expressing its attention, as well as basic emotions, such as pleasure and excitement, it is able to enhance the social abilities of ASD children (Kozima et al.,2007).

Probably the most famous example within ASD research is the robot called KASPAR (see Figure??). It was developed explicitly for the purpose of supporting children with ASD and according to pro- fessionals, it yields a high potential for a broad range of therapies and educational goals in this scope (Huijnen et al.,2016). In fact, different studies have already shown that KASPAR can support children suffering from ASD to develop social and communicative skills, as well as to explore and share their basic emotions (Robins et al.,2012;Wainer et al.,2014). Furthermore, the developed skills were shown to last over a longer period of time, even when the robot is gone again (Robins et al.,2005;Wainer et al.,2014). But, the development of SARs applicable in this field is still ongoing and new concepts are under development, e.g., to relieve therapists or experimenters by increasing the robot’s autonomy during the interaction (Zaraki et al.,2018). In addition, concepts for enriching the diversity of robots in ASD therapy can be found, which focus on robots that are already frequently applied in other fields, such as Nao (So et al.,2019;Yang et al.,2019).