JHRI Template - ACM Digital Library

Unconventional Students in HRI Education: A Review of Two Initiatives Julaine Zenk,1 Charles R. Crowell,1 Michael Villano,1 Juhi Kaboski,1 Karen Tang,1 and Joshua John Diehl1,2 1 Department of Psychology, University of Notre Dame 2 LOGAN Community Resources, Inc., South Bend, IN Robotics and human-robot interaction (HRI) are growing fields that may benefit from an expanded perspective stimulated by more interdisciplinary contributions. One way to achieve this goal is to attract non-traditional students from the social sciences and humanities into these fields. This present paper describes two educational initiatives that focused on teaching non-engineering students about robotics and HRI. In one initiative, a group of younger students, including those with autism spectrum disorder (ASD), received hands-on experience with robotics in a context that was not overly technical, while in the other initiative, college students in the social sciences and humanities learned about basic HRI concepts and developed robotics applications. Themes common to both initiatives were to reach non-technical students who are not traditional targets for robotics education and to focus their learning on creating interactive sequences for robots based on key HRI design considerations rather than on the underlying mechanical and electrical details related to how those sequences are enacted inside the robot. Both initiatives were successful in terms of producing desired learning outcomes and fostering participant enjoyment. Keywords: robotics, HRI, education, interest-driven exploration, students with ASD, non-traditional robotics students

Introduction As robots become more prevalent in society, an understanding of human-robot interaction (HRI) becomes more relevant and the need for robotics-related education becomes more pressing. This present paper describes two educational initiatives that took place at the University of Notre Dame, both of which focused on teaching nonengineering students about robots and HRI. In one initiative, adolescent students, some with autism spectrum disorder (ASD), received hands-on experience with robotics in a context that was not overly technical, while in the other initiative, college students in the social sciences and humanities learned about basic HRI concepts and developed robotics applications. Goals common to both initiatives were to first reach non-technical students who are not traditional targets for robotics education; second was to focus their learning on creating interactive, social sequences for robots rather than on the underlying mechanical and electrical details. We believe there is great merit in drawing non-engineering students into the fields of robotics and HRI. Ultimately, we think this strategy will enhance research and innovation in both fields by anticipating technological innovations (Diehl, Schmitt, Villano, & Crowell, 2012). However, several potential barriers can prevent or diminish interest among non-technical students who might otherwise pursue these fields. These barriers include the following: pre-conceptions related to potential career opportunities in the field of robotics, the lack of pre-existing interests in science and technology among students, the lack of understanding of the broader relevance of robotics today, the inaccessibility of the field of robotics to prospective students, and the challenge to better capitalize on the interest in

Authors retain copyright and grant the Journal of Human-Robot Interaction right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal. Journal of Human-Robot Interaction, Vol. 6, No. 2, 2017, Pages 92-110, DOI 10.5898/JHRI.6.2.Zenk

Zenk et al., Unconventional Students in HRI Education

technology and robotics that may accompany special clinical populations, such as individuals with ASD. All of these barriers, including the special social challenges confronting individuals with ASD, have guided our thinking about the kinds of educational experiences necessary to draw non-technical students into the fields of robotics and HRI.

Early Perceptions of Robotics May Be Limited While in recent years more engineering and technology courses are being offered at the grade school level, they are still largely unavailable to students at the pre-secondary level (Riedo, Chevalier, Magnenat, & Mondada, 2013). Therefore, young students may have limited knowledge of the various ways robots can be useful in modern society, as well as a pre-conception about what kind of background is required for robot application development. A sample of students in fourth through sixth grade (24 boys and 24 girls) was surveyed regarding perceptions of robotics using three response categories: “robots as a plaything… learning of robotics as a source of employment… [and] learning of robotics as a way to high technology” (Liu, 2010, p. E45; Tsai, 2004). In this survey, 87.5% of the students reported that they regarded robots as playthings, while only 50% reported that learning about robotics could be a possible source of employment (Liu, 2010, p. E45). Most of those who regarded robotics as a possible employment opportunity had in mind the mechanical aspects of robots or manufacturing uses; expressing an interest in pursuing an engineering degree. Liu (2010) also pointed out that a significantly larger number of male students cited an interest in robotics employment compared to female students (t = 4.108, p < 0.001; p. E46). In our work, we observed gender-related differences in how robots are perceived (Crowell, Shermerhorn, Scheutz, & Villano, 2009). If this gender disparity trend continues, it should be addressed in specific ways involving both robotics design and education. For example, Liu (2010) suggested that robots with more appeal to females could be created, which is consistent with the notion Crowell et al. (2009) espoused. Gender-related characteristics (e.g., appearance, voice) of robots may affect how people identify with them along gender lines (Crowell et al., 2009). Also, Liu (2010) suggested that many female students who might be interested in the field of robotics may not have the confidence to pursue study in this area, as this field has not been well marketed toward the female demographic. Improved robotics education and marketing is needed to increase the number of females working in the field of robotics.

Cultivating Pre-Existing Student Interests in Science and Technology Many camps, clubs, and competitions related to robots have provided relevant learning opportunities to draw more students into these fields. These opportunities capitalize, at least in part, on broader, existing interests students already have in science and technology (Nugent, Barker, Grandgenett, & Welch, 2016; Kaboski et al., 2015). After eight years of developing and studying the outcomes of these types of programs, Nugent et al. (2016) noted that students with personal interests in science and technology were more likely to get involved in camps, clubs, and competitions when they found ways to develop interesting and relevant hands-on robotics applications. Sparking personal interests through robotics camps and clubs gets students more involved in relevant science and technology learning exercises. Furthermore, it affords students with opportunities to connect with fellow campers, facilitating social connection and collaboration skills. Kandlhofer and Steinbauer (2016) examined the effect of robotics education on technical and social skills. Their program took place over multiple months and involved non-engineering, adolescent students (M = 14.9 years) with no robotics experience participating in the RoboCupJunior (RCJ) competition. In Kandlhofer and Steinbauer’s (2016) study, half of the students were placed into an experimental group and were enrolled in “robotics activities” centered on preparing for a national RCJ competition. A control group participated in computer science courses that taught students about current application software, how to create websites, and also involved online research to present reports using new media resources (e.g., tablets, PowerPoint, email; Kandlhofer & Steinbauer, 2016). RCJ was described as an “international educational robotics initiative, aiming to promote STEM (science, technology, engineering, and mathematics) content and skill learning” (Eguchi, 2015, p. 692). In an effort to attract more students, especially females, to the program, contestants were encouraged to incorporate their own interests into their robotics projects. Kandlhofer and Steinbauer (2016) reported a relatively large percentage of female participants (40%) in the RCJ experimental group, which could have been due to this interest-driven approach (p. 682). 93


Based on pre- and post-test comparisons, Kandlhofer and Steinbauer (2016) indicated that the experimental group showed significantly larger increases in technical and social skills than the control group. Additionally, robotics students in the experimental group exhibited more positive attitudes about science and teamwork (Kandlhofer & Steinbauer, 2016). Similar outcomes were obtained in a related study by Eguchi (2015), who observed improved knowledge of STEM topics overall and improved creative skills. Taken together, these findings suggest that teaching non-engineering students robotics nurtures their interest in technology and science while increasing collaboration and creativity. It is noteworthy that significant improvements have been reported in educational programs that involve technology-related, hands-on guided inquiry (Kurth & Mastergeorge, 2010). Through guided inquiry, students use the advice and direction provided by teachers to assume the role of scientists, seeking out answers to their own interest-driven questions. This approach has been used to increase children’s inferential reasoning and social communication skills (Kurth & Mastergeorge, 2010). Following suit in robotics education by encouraging individuals to incorporate their own interests into projects and applications could yield strong benefits for the field. This approach might broaden the range of disciplines that contribute to robotics and HRI, which could facilitate further innovation in these fields. Moreover, encouraging young students to incorporate their individual interests into the study of robotics and HRI applications might well result in more diverse, interdisciplinary collaborations within these fields.

Capitalizing on Technology/Robotics Interests Among Students With ASD The clinical population of students with ASD is a veritable wellspring of individuals with a known interest in science and robotics (Kim et al., 2013). Facilitating interest-centered activities related to robots among these individuals is a useful technique employed when working with children with ASD in both educational and therapeutic environments (Bak & Siperstein, 1987; Dunst, Trivette, & Hamby, 2012). Several investigators have observed that interactions with robots motivate prosocial behaviors and learning in children with ASD (Dautenhahn, 2003; Diehl et al., 2012; Kim et al., 2013). Moreover, it has been noted that the prosocial behaviors gained during interactions with therapeutic robots transfer to real-life interactions with peers and adults (Boyd, Conroy, Mancil, Nakao, & Alter, 2007; Diehl et al., 2012). Perhaps, by drawing on their interest in robots, children with ASD can be motivated and encouraged to collaborate with their typically-developing peers on robot development projects (Kaboski et al., 2015; Wainer, Ferrari, Dautenhahn, & Robins, 2010). Other research leads to similar conclusions. For example, Wainer et al. (2010) noted an improvement in sociocognitive behavior in students with ASD within the context of a collaborative robotics programming project. Bauminger-Zviely, Eden, Zancanaro, Weiss, and Gal (2013) conducted a similar study for participants with ASD using a non-robot-based, collaborative technology intervention. These researchers reported that their intervention resulted in high levels of task completion and satisfaction among their participants with ASD and also increased participants’ motivation for socialization with typically developing peers. Cabibihan, Javed, Ang, and Aljunied (2013) found that children with ASD showed improvement and enjoyment when they were able to make choices during their interactions with a robot. Similar observations were reported by Kaboski et al. (2015) regarding the Summer Robotics Camp initiative described later in this paper.

Expanding the Perceived Reach of Robotics As noted earlier, Liu’s (2010) survey of youth revealed that many respondents had a limited view of the relevance and applicability of robotics in modern society. It is likely that such views result from the prevalence of robotic applications in areas such as manufacturing, where the goal of the robot is to automate processes that otherwise would require human actors. To broaden perspectives on robotics, it is necessary to expose students to other possible uses and applications of robotics in areas beyond task automation. Social robotics is one such area in which there exists a vast array of opportunities to exploit the utility of robots for providing humans with companionship, assistance, entertainment, or information. Recognizing and exploring these types of applications will broaden a student’s horizons regarding the relevance and reach of robotics. 94


One initiative along these lines was RoboWaiter: The Assistive Mobile Robot Competition. This program was designed to foster projects devoted to social, moral, and humane issues related to assistive technology (Ahlgren & Verner, 2013). Participants in this competition were high school students and first-year prospective engineering undergraduates. Representatives from the Connecticut Council on Developmental Disabilities provided guidance and feedback to the competitors. The competition also involved focus groups consisting of individuals with disabilities to help inform the judges’ decisions. All of this served to underscore the importance of creating applications that would have value from the viewpoints of industry professionals, as well as end-users with disabilities. Overall, the RoboWaiter competition accomplished two important goals. First, it broadened the perspective of participants by raising their awareness of the salience of robotics in assistive technologies. In a post-contest survey, many participants reported a better appreciation for the humanity of engineering and a new passion for the use of robotics in assistive technologies. One participant noted, “As an engineer, I tend to get focused on a task and forget about the people... For an assistive robotics project, I think it would be very valuable for students to interact with people who have disabilities and get their feedback” (Ahlgren & Verner, 2013, p. 134). Second, RoboWaiter allowed participants to develop a better appreciation for the benefits of cross-field collaboration. Similar goals have been put into practice under the auspices of the Open Platform for Social Robots (OPSORO) emerging from the University of Gent in Belgium (Vandevelde & Saldien, 2016). OPSORO is an open platform for creating social robots that focuses on face-to-face communication and creates a bridge between social scientists, therapists, and collaborators from engineering fields (Vandevelde & Saldien, 2016). This initiative aims to expand the perspectives involved in creating functional robots by including greater awareness of what aspects are needed for a specific therapy, information that may be best obtained from practicing therapists. OPSORO essentially allows socialscientist-therapist teams to become roboticists by replacing the complex aspects of robot design with a modular approach to embodiment creation. The robot creator uses these easy-to-understand design modules to generate a robot skeleton and internal infrastructure capable of completing the necessary tasks while also having an appropriate outward appearance. Once designed, the actual robot is constructed through laser cutting or 3D printing technologies. After construction, the robot can be programmed through a browser-based control software interface implemented over Wi-Fi. Through programs, such as RoboWaiter and OPSORO, non-engineering audiences are better able to design and implement robotic platforms by focusing more on the social and human issues to be addressed rather than the technical aspects of robot design, construction, and control. In this way, it may be possible to create more socially assistive and effective robot applications. Future robotics competitions should take advantage of these developments to empower and jumpstart non-technical interdisciplinary teams as they aspire to build the socially-assistive robotic technologies of the future.

Enhancing the Accessibility of Robotics The field of robotics can be intimidating for many reasons. For some, especially younger students, a robot’s appearance and/or behavior can determine attraction or repulsion. If a robot appears menacing, many people (and most children) may not want to approach it (Gorostiza & Salichs, 2011). The well-known ‘uncanny valley’ hypothesis suggests that the more human-like a robot is in appearance and behavior, the more it is perceived as approachable (i.e., positive responsivity of viewers) up to a point (Mori, 1970). According to this hypothesis, if a robot looks almost human but doesn’t quite behave that way, it can become eerie or even repulsive to its viewers. In contrast, when a robot looks friendly, approachability may be enhanced (Gorostiza & Salichs, 2011). This is especially true for younger viewers or those in special clinical populations. For example, in a recent review of the clinical uses of robots in therapy for ASD (Diehl et al., 2012), it was noted that children were more relaxed and exhibited lower heart rates when interacting with attractive robots. Also, this review indicated that children with ASD preferred robots that looked less like a human and were more character-like or robotic (Diehl et al., 2012). The technical aspects of how to program a robot and make it interactive can also be a daunting barrier to those students with robotics interest and no technical experience (Gorostiza & Salichs, 2011). However, recent developments have reduced the entry-level technical background needed by those who aspire to develop interactive sequences for robots (Gorostiza & Salichs, 2011; Magnenat, Riedo, Bonani, & Mondada, 2012; Riedo et al., 2013). 95


For example, social robots, such as Maggie and Thymio II, were designed to be aesthetically attractive and easily controlled. Maggie uses a Natural Programming System (NPS) that allows end-users to create interactive sequences through voice commands rather than complicated programming methods (Gorostiza & Salichs, 2011). Thymio II uses open-source software, called Aseba, which is designed specifically for novice programmers utilizing a Visual Programming Language (VPL). NPS employs a human-centered approach to teaching programming and has been observed to be less error prone for new programmers. With Maggie, the NPS system uses natural voice commands from the programmer, which the robot decodes and filters using special onboard VoiceXML software (Gorostiza & Salichs, 2011). The VPL employed by Thymio II uses cards representing events and actions strung together in a timeline that simplifies writing real-time behaviors (Magnenat et al., 2012). Thymio II’s programming also employs built-in behaviors and conditions that give beginning programmers an easy starting point for projects. Thymio II also has extensions to Blockly, Text Programming, and Scratch 2 scripting languages, which allow novice programmers to develop and practice advanced skills once they are comfortable with the basic functions offered by the VPL (Riedo et al., 2013). NPS and VPL programming represent innovative approaches to the teaching of robotics, which is particularly useful for programming social robots, since it permits a user to program, test, and re-test with ease (Gorostiza & Salichs, 2011; Magnenat et al., 2012; Riedo et al., 2013). Reducing the entry-barrier through these kinds of coding techniques will likely encourage discovery learning and will allow those without pre-existing science and technology experience to get involved in robotics application development without the steep learning curve required in developing traditional programming skills. The educational initiatives reported here employ a similarly user-friendly programming interface that reduces the entry-barrier. In both initiatives, robot control was achieved using Choregraphe (Aldebaran, 2016), a visual, flow-chart based interface that is easy to learn, thereby precluding the need to be familiar with traditional programming syntax or other coding languages. The control flow charts in this system are like those in Thymio II’s VPL system and consist of behavior boxes (e.g., speech, preprogrammed movements) that are connected via lines that the user creates to construct a timeline for their overall project. More in-depth and complex programming can be done with logic boxes (e.g., conditionals or switch statements) or within each behavior box, although these topics were not covered in the Summer Robotics Camp (first initiative described below) due to time constraints and minimal prior robotics knowledge. The Practicum in Robotics (second initiative described below) students, however, were encouraged to consider these capabilities if they chose. The Choregraphe software platform is provided with the Aldebaran Robotics NAO Academics Edition robot. This robot is a 23-inch tall humanoid robot that is capable of online text-to-speech communication and 25 degrees of freedom in movement, which allows for convincing human-like social gestures.

Overcoming Barriers to the Study of Robotics and HRI In sum, research has confirmed entry-level barriers to the fields of robotics and HRI among many non-traditional students. Such students may approach the field of robotics and HRI with a very limited knowledge base (Liu, 2010). Novice students may only view robots as “playthings” instead of considering academic or employment possibilities in the field of robotics. Furthermore, there is an overwhelming lack of interest in robotics and its applications among females and non-technical students, which needs to be addressed (Ahlgren & Verner, 2013; Crowell et al., 2009; Liu, 2010). Additionally, perceived requirements for further study in these fields may seem daunting to many. In response to these barriers, more programs and initiatives need to be created to cultivate pre-existing interests in science and technology, to expand the perceptions of robotics capabilities, and to enhance entry-level accessibility to the fields of robotics and HRI. Accessibility is a key aspect of reaching those students who are not science and engineering oriented and who are inexperienced with robotics. Through programs such as RCJ, participants and attendees can see the wide array of possible robot technology uses and HRI applications (Eguchi, 2015; Kandlhofer & Steinbauer, 2016). Furthermore, the collaborative focus of the RoboWaiter competition highlights the importance of interdisciplinary cooperation for enhancing innovation and research (Ahlgren & Verner, 2013). Other programs also involve a guided inquiry learning style that helps to motivate students by allowing the role assumption of a scientist and incorporating unique 96


interests into the students’ robotics learning (Kurth & Mastergeorge, 2010). Through these kinds of programs, a student’s innate interest in robotics and technology can be fostered. As we noted above, barriers to entering the field of robotics can be diminished in several ways. First, robots can be made easier to use through methods such as NPS programming (Gorostiza & Salichs, 2011) and VPL (Riedo et al., 2013), allowing novices to start programming without the entry cost of extensive technical backgrounds. Second, accompanying a reduced emphasis on the technical aspects of programming and control, greater focus can be placed on creating desired and relevant robot behaviors instead of mechanics and technical programming, thus creating a more accessible learning experience (Ahlgren & Verner, 2013). Finally, robots can be made more approachable and attractive (Diehl et al., 2012; Gorostiza & Salichs, 2011; Mori, 1970) through initiatives like OPSORO (Vandevelde & Saldien, 2016). With these steps, the field of robotics can become a more attractive option for many students and academics who want to include robotics in their studies. These steps will enable a shift in focus to the human and assistive issues that properly belong to HRI, potentially opening the door to younger students, including those with certain developmental disabilities such as ASD.

Our Educational Initiatives Many, if not all, of the aforementioned barriers for non-technical students have been addressed by the two robotics educational initiatives presented below. The student populations in both initiatives came from the local community or from social science and humanities departments at the University of Notre Dame. All participants came to these initiatives based on emerging interests in the topic of robotics. Both initiatives involved hands-on, behavior-based programming methods, making these educational experiences more accessible for the non-technical participants. The broad goal of both initiatives was to nurture student interest in the application of robotics technology to everyday life and to the associated HRI issues that arise in such applications, while also encouraging teamwork and collaboration. A secondary goal was to potentially motivate further study and involvement in the field for at least some participants.

Relevant HRI Considerations Certain relevant HRI considerations were common to both of the educational initiatives to be described here. Over the years, the HRI community has devoted some attention to the matter of identifying important design and interaction principles to guide HRI development (e.g., Goodrich & Olsen, 2003; Forlizzi, DiSalvo, & Gemperle, 2004; Kahn et al., 2008; Kiesler, 2005). While a universally accepted set of HRI guidelines has yet to emerge, we have identified several categories of questions that serve as relevant design considerations in guiding our own HRI work and that of our students. We incorporated some or all these questions in both of the educational programs described here. In the sections below devoted to each specific initiative, we describe how we utilized these guiding questions. We encourage our students to think about at least five broad categories of questions when designing applications that involve human-robot interactions. These categories may not be the only relevant considerations, and each of these questions may not always apply. However, in our experience, these questions offer a useful framework for thinking about specific HRI situations. 





Audience/end-user (AEU) knowledge, experience and expectations - What does the AEU already know about the situation in which the interactions will occur? - What does the AEU already know about the robot with whom they will interact? - What expectations does the AEU have about what will/should happen in this situation? AEU goals - What are the needs of the AEU in this situation? - What does the AEU want to get accomplished? - What does the AEU want to avoid? The robot platform to be employed - What can it do? - How can it sense? - How will its appearance/presence be received by the AEU? 97




The human-robot interactions in the situation - How will the AEU communicate with the robot? - How will the robot communicate with the AEU? - Under what circumstances and how often will these communications occur?



Emotional cues - Are emotional cues from the AEU relevant? If so, how will they be detected? - Are emotional cues from the robot relevant? If so, how will they be conveyed?

Initiative 1: A Summer Robotics Camp A Summer Robotics Camp was developed in collaboration with personnel at the William J. Shaw Center for Children and Families at the University of Notre Dame. The camp was geared toward middle and high school students, ages 12 to 17, half of whom were diagnosed with ASD. The idea for the first Summer Camp emerged from observations made during a therapy program being conducted at that time involving individual children with ASD, a human therapist, and a robotic co-therapist (Diehl et al., 2013). During the therapy, we noticed that older individuals with ASD were just as drawn to learning how the robot worked as they were to interacting with it during the therapy sessions. This observation led directly to the development of a camp experience in which therapeutic (social and career skills) goals could be combined with learning about robots in an innovative educational program. A major component of the learning opportunity we envisioned for the camp would involve having the participants develop their own robot stimulus-action sequences, which they would try out on themselves first to see how the robot responded, making any necessary adjustments before making their sequences available to others. To set the stage for participant understanding of HRI, the camp started with a general session in which many of the guiding questions described above were discussed. Special emphasis was placed on factors related to the appearance and appeal of the robot, as well as the potential roles robots could take in a human society. The Summer Robotics Camp was conducted over two consecutive summers using essentially the same procedures each time. A detailed published report on some of the findings from the first summer already has been provided by Kaboski et al. (2015). A report on the second summer camp is now under development. The camp information presented here will be limited to a description of the procedures employed, along with selected data from the first and second summer camps. Also, we include a new analysis of some data from the first camp. The forthcoming follow-up report will detail results and outcomes from the second camp. Even though there is a previously published report on the first camp (Kaboski et al., 2015), we included this initiative in the present paper because it is an innovative HRI educational program, including both typically developing adolescents as well as those with ASD, that is highly relevant to the broader goal addressed in this paper of making robotics and HRI education available to non-traditional populations. In addition, we have at least one new finding to report here about the first camp. Camp Goals The educational goals of the weeklong camp were to learn robotics facts, to program an interactive robot, and to employ vocational skills. Additionally, the camp’s goals were to make the camp experience available to a population of children with ASD who otherwise would not be able to partake in these kinds of educational opportunities due to their social limitations. Camp Organization Participants were recruited from the local community and from the existing listserv of local families who had participated in or indicated an interest in other ASD-related research studies. Inclusion criteria were: (1) a selfreported interest in robotics, (2) an absence of any untreated psychiatric, developmental, or physical diagnoses that could interfere with participation in the study, and (3) enrollment in general education science classes during the academic year (with or without classroom accommodations, such as a one-on-one aide). The final sample consisted

98


of sixteen boys, half with ASD and half typically developing (TD), within the age range of 12 to 17 years. Although we did not intentionally exclude females in our sampling procedure, the final sample consisted of exclusively male participants. After screening and selection, participants were paired into dyads, matched as closely as possible on age, previous robotics experience, IQ, and language skills. Each dyad consisted of a TD adolescent and an adolescent with ASD. The weeklong camp consisted of five, three-hour days. The first four days included instruction and hands-on experience with Choregraphe and the robot, while the fifth day allowed students to complete and present final robotics projects to an audience. The students began each of the first four days with group instruction covering a technical robotics programming skill and what were dubbed “career skills.” The career skills lessons consisted of strategies for real-world science collaboration. Technical skills and career skills were interrelated each day. For example, on the day the participants learned about voice recognition and face tracking, they also learned the career skill of listening and the benefits of communicating clearly with others. After instruction, students worked in their assigned pairs developing robot control sequences using the skills learned that day. During programming skills practice, two camp facilitators monitored and assisted campers. These facilitators also led large group instruction. On the fourth day, instruction and programming were geared toward constructing and completing a final robotics project. Pairs had to work together to decide on a project that was interactive and based on a shared interest. After a theme was selected, student pairs developed a control sequence that would allow the robot to enact what would be presented to an audience on the last day of camp. For example, one pair chose to program the robot to tell jokes. They programmed the robot to introduce herself, ask a question, acknowledge the answer from the audience, and respond in a socially appropriate manner while simultaneously using appropriate hand gestures. By working together on interactive projects of this sort, the team pairs not only worked on their robotics skills, but on their social skills as well. One obvious social skill was teamwork, but another of great importance was based on the fundamental HRI consideration (noted in the above questions) of how the robot’s behavior could be designed for maximum audience impact given the specific goal of the team’s chosen demonstration. Interestingly, the latter consideration often forced both team members to reflect on their own behavior (i.e., how they would act if they were the robot), which often was very beneficial and instructive for both team members, but especially for the member with ASD. For example, the camp participants were explicitly reminded that, in an effective human-to-human interaction, the speaker should pay attention to his audience’s interests and level of understanding. Subsequently, he also should adjust his speaking tone, volume, and vocabulary depending on the recipient of the communication, whether it is a large group, a small child, or an adult. Therefore, in order to facilitate the most natural and useful interaction between the robot and its communication partner, the campers needed to design the robot to first recognize “who” the robot was addressing and then choose the most appropriate communication mode. This approach was an attempt to demonstrate, in a tangible yet non-threatening way, the functions behind why humans practice certain social/communication skills, as well as to encourage participants to become mindful of what robots need to do in order to appear more human-like. For children with ASD who may have never understood the relevance behind certain social skills or never felt compelled to practice what they knew, this gentle reminder in the context of HRI could be used as motivation for the students to practice the same social skills in their everyday interaction with others, a hypothesis that still needs to be tested. On the last day of camp, a reception was held for the campers, camp facilitators, family, and friends, during which the campers presented their projects to a large group. The crowd was also encouraged to ask the participants questions about their project or about the camp as a whole. This was an opportunity for the campers to showcase not only the technical skills they had learned throughout the camp, but also to utilize some of the new career skills they learned. For example, rather than introducing oneself, each participant had to introduce his partner to the audience. They were encouraged to make appropriate eye contact and natural gestures during presentation, and explain in coherent and socially engaging language the reasons behind their choices or the ways in which they collaborated with their teammate. At the completion of the second camp, participant feedback was solicited in order to gauge their overall subjective experience and to incorporate their suggestions for future improvement of the program. Campers also completed a robotics knowledge quiz that covered materials taught during the camp. These results were compared to an identical quiz given before commencement of the camp. 99


Relevant Camp Educational and Behavioral Outcomes As reported by Kaboski et al. (2015), campers were tested on social anxiety, social skills, and knowledge of robots and robotics, both before and after the weeklong camp. For additional information on the measures used and statistical results on these measures, please refer to Kaboski et al. (2015). An abbreviated summary of those results can be found in Table 1. Table 1. Pre-intervention vs. post-intervention data from Kaboski et al. (2015) Pre-intervention M (SD)

Post-intervention M (SD)

t

p

d

ASD Group (N=8) SAS-A/SASC SSIS Robotics Knowledge Quiz

43.38 (9.15) 74.13 (16.49) 2.17 (1.52)

37.38 (6.82) 79.38 (14.46) 7.98 (1.52)

2.89 -1.79 -13.03

.02* .12 < .001**

.74 .17 2.73

TD Group (N=8) SAS-A/SASC-R SSIS Robotics Knowledge Quiz

32.63 (10.43) 109.75 (8.71) 1.42 (1.30)

32.75 (9.35) 109.25 (11.23) 5.15 (1.73)

-.12 .20 -7.16

.91 .85 < .001**

.01 .05 2.47

Measures

Note. SASC-R = Social Anxiety Scale for Children – Revised; SAS-A = Social Anxiety Scale Adolescent; SSIS = Social Skills Improvement System. *p < .05, **p < .01. Data taken from “Brief Report: A Pilot Summer Robotics Camp to Reduce Social Anxiety and Improve Social/Vocational Skills in Adolescents with ASD,” Journal of Autism and Developmental Disorders, 45, p. 3866, Copyright 2014 by Springer Science+Business Media. Adapted with permission.

As the above table shows, campers experienced a significant decrease in social anxiety at the completion of the camp. Before the intervention, the average social anxiety level as measured by Social Anxiety Scale (SAS; La Greca & Stone, 1993) was 43.38 (SD = 9.15) for the ASD group and 32.63 (SD = 10.43) for the TD group. The ASD group showed a significant reduction in social anxiety from the baseline to post-intervention (p < .05) and became statistically indistinguishable from their TD peers (p = .28). The average baseline level of ASD participants’ social skills using the Social Skills Improvement System (SSIS; Gresham & Elliott, 2008) was 74.13 (SD = 16.49), 1.5 standard deviations lower than the general population mean of 100. The TD group scored 109.75 (SD = 8.71), consistent with the general population mean. As Kaboski et al. (2015) reported, the mean SSIS score did not change significantly from pre- to post-test for either group, though the improvement within the ASD group approached significance. For this paper, we conducted a new analysis of the social skills results reported by Kaboski et al. (2015) based on the reported finding that six of the eight participants with ASD showed an arithmetic increase (i.e., improvement) in their overall SSIS scores from the pre- to post-test, even though the mean change in mean SSIS scores was not quite significant for this group. The number of ASD participants showing pre-post improvement in SSIS scores contrasted sharply with the number of TD participants showing comparable change in that only three of the eight TD participants showed an arithmetic increase. To examine this difference in the number of participants who changed more closely, we computed pre-post difference scores on the SSIS measure for each participant in both groups. The average change for the TD group from pre- to post was -0.75, while the comparable change for the ASD group was +5.25. We applied a median test to these difference scores using the Fisher Exact procedure, which revealed that the number of subjects showing a change in excess of the overall (across groups) median change (+2.5) was significantly greater in the ASD group (6 participants) than in the TD group (2 participants; Pearson’s chi-square = 4, p < .05). Thus, in terms of outcomes resulting from the camp, participants with ASD exhibited a significant decrease in social anxiety scores from pre- to post-tests, and significantly more ASD participants than TD 100


participants showed an improvement in social skills scores. These two areas, social anxiety and social skills, are hurdles that often prevent individuals with ASD from participating in or fully benefiting from education programs or camps, which is why we targeted these measures in this initiative. Additionally, as Kaboski et al. (2015) noted (see Table 1), all 16 participants showed significant pre-post increases in robotics knowledge exam scores, an outcome that did not differ between ASD and non-ASD campers. This finding means that all participants learned something about robotics as a result of their camp experience, an outcome fully consistent with the above-noted camp goals. Exit interview feedback. The exit interview was an open-ended survey that was conducted with the 36 participants from the second year of the summer camp. The exit interview data reflect directly upon the experiences of camp participants in relation to the above-noted camp goals, which were the same for both iterations of our two camps. The exit interview reported here was not conducted for the first summer camp; however, based on unsolicited comments from first camp participants and their parents, we believe the interview results below also characterized the experiences of participants from the first camp. The exit survey included questions such as, “What was your favorite part of the camp?”; “What was it like working with a partner?”; “What was easy?”; and “What was hard?” To summarize the responses, a coding scheme was generated for each question. Two researchers coded the campers’ responses independently with a mean Cohen’s Kappa of 0.81 for the inter-rater reliability. A Kappa above 0.80 is considered a high level of agreement (Cohen, 1960). When asked about their favorite part of camp, 62% of the campers noted that they most enjoyed the portion the camp devoted to learning how to control the robot. Also, many campers cited their enjoyment of this component with statements such as, “Being able to program an awesome robot” and “My favorite part was just working with Lisa [name given to robot] and doing cool things with her.” The campers were also asked what they believed was the most important thing they learned during camp, and about half (48%) said they thought the control sequence element was most beneficial to them. Teamwork and social skills were also frequently cited as most important (28%). Campers were also asked specifically what it was like to work with a partner. A little over half (53%) of the participants reported that they had no difficulty working with their partner; 25% said they had mild to moderate difficulty but were able to work through any issues. For example, one participant said: “At first it was hard because I had never met my partner, but once I got to know him, it got easier.” Twenty-five percent of campers stated that they struggled at working with their partner: “It was difficult to coordinate our ideas and to determine what would work and what wouldn’t” and “It was kind of hard. I didn’t get to know him very well.” The two cited statements are indicative of many statements made by the campers who struggled. Many struggled to coordinate their efforts, as they did not feel they knew their partner well enough to encourage their partner to contribute more or to assert themselves. This could be easily related to the fact that the camp only spanned five, three-hour days. The final question of the exit interview asked, “If you were to do the program again, what else would you want to learn?” Many of the participants (33%) noted that they would want to learn more about the capabilities of the NAO robot, such as interactive capabilities, more behavior boxes, and general functionality. In addition, many the campers (53%) noted that they would like to learn more technical details about how the control boxes worked and what was going on inside the robot. Initiative 1 Conclusions In this first educational initiative, a weeklong Summer Robotics Camp was created to increase participant knowledge about robotics, provide a hands-on opportunity to control a robot, and incorporate HRI concepts, vocational, and collaboration skills. Also, it was hoped that the camp would nurture social skills and reduce social anxiety in participants with ASD because of their partnership with a TD peer. As Kaboski et al. (2015) reported, and as the exit interview responses presented above reflect (if they can be taken as indicative of first camp participants), the objectives of the first camp were met. Knowledge of robotics increased across the entire group, which we believe was largely due to the structured learning sessions and the hands-on teaching methods employed in the camp. Throughout camp, the pairs of campers oversaw their own

101


assignments and, as pairs, collaborated to decide what theme to focus on in their final projects. Thus, each pair could draw on their shared interests to create an interactive presentation with the NAO robot. Most campers commented that they learned the most about controlling and creating interactive sequences using the VPL software, Choregraphe. Furthermore, a majority of these campers noted that it was easy to work with their partners when creating their final projects. Social anxiety was reduced significantly for the campers with ASD, and many individual participants with ASD saw improvements in social skills as well. Further, the involvement of HRI and interactivity in their projects helped to solidify social skills learned throughout camp. As a result of the hands-on learning and the interest-based design of their final projects, many campers reported they were motivated to learn more about the field of robotics and HRI. While some participants already had interests in these topics, their knowledge and interests were sharpened and refined by the camp experience. Participants were better able to see the potential applications within the field and how they could apply their interests in a vocational sense.

Initiative 2: A Practicum in Robotics for Non-Engineering College Students Our second educational initiative was an undergraduate course entitled “Practicum in Robotics.” This course is offered through the University of Notre Dame’s Computing and Digital Technologies (CDT) minor, a program intended for students in the liberal arts arena. CDT gives liberal arts students the opportunity to add an interdisciplinary aspect to their undergraduate education by blending “programming and technology skills with the liberal arts in a wide variety of ways” (University of Notre Dame, 2016). Practicum in Robotics is a semester-long (14 week), project-based course that is self-exploratory in nature. Students taking the course worked with the same NAO robot platform used in the Summer Robotics Camp described above. Course Goals This course allows students to work with the NAO humanoid robot platform in order to learn about its various capabilities and gain experience in developing applications involving NAO-human interactions. Specific student activities and learning goals in the class included:   

 



Discovering how to control the robot to produce specific sequences of robot behaviors and/or to allow the robot to respond to particular inputs from its external environment. Using instructor lectures and examples, along with online learning resources and tutorials made available by the instructors, to guide self-exploration of NAO platform capabilities. Completing specific weekly assignments in the first half of the course that involved tasks such as developing specific behavioral sequences for the robot, experimenting with the robot’s sensory abilities, and exploring the various methods available to program and control the robot. Learning about and considering the applicability of five important categories of HRI considerations that should guide the development of human-robot interactions. Identifying and then developing a final project to be presented at the National Robotics Week (NRW) Exhibition on campus in April that conforms to the five categories of HRI considerations, given a general audience consisting of children and adolescents, ages 4–16 years, along with their parents. Attending the NRW exhibition (since its inception in 2012) and demonstrating the final projects to the audience that visited the lab’s exhibition booth while also answering questions about the robot’s capabilities, methods of programming, and its capabilities for HRI.

Course Organization The practicum typically has had a small enrollment in each semester. Although the class antedated the initiation of the NRW exhibition at the university, over the years since the exhibit began in 2012, the average class size has been 102


4 students per semester. A total of 19 students have been enrolled over that time, including one semester with no enrollment. Of those enrolled, 13 were male and 6 were female. Sophomores through seniors were eligible to enroll, which produced an approximate age range from 18–22 years. The format of the practicum provided students a great deal of flexibility in how they went about exploring the NAO platform hardware and its control development software and capabilities. Throughout the semester, students used self-guided inquiry to learn about the various capabilities of the NAO robot. Over the course of the first seven weeks, students completed weekly assignments to familiarize themselves with Choregraphe and the NAO platform. Assignments included the following:       

Create simple behaviors—at least one limb movement and one vocalization in sequence or in parallel. Create more complex behaviors—multiple movements of different limbs combined with vocalizations and (eye) LED changes. Create stimulus-response sequences—behavioral sequences triggered by a sensory input set up as a recurring loop. Use movement editor and timeline—create and edit a movement adjusting timing and joint positions. Programming with Choregraphe—using different kinds of box inputs and outputs to achieve some logical control over behaviors (e.g., timing, counting, or randomization). Advanced NAO programming—implementing behavioral sequences using external programming methods or by changing the features of a Choreographe box. Explore the use of DOMER, an application developed for Wizard of Oz control of the NAO robot (Villano et al., 2011). DOMER was initially created to facilitate ABA-style therapies between the NAO robot and children with ASD.

After completing the first seven weeks of structured assignments, students continued their self-guided exploration of the NAO platform, but also began to focus attention on their final projects. Final projects were defined by students in consultation with course instructors. An important part of defining this project involved the five categories of HRI considerations noted above. Final projects needed to take into consideration the various questions associated with each of the five HRI design categories that were appropriate for the NRW audience. Students needed to consider and address relevant HRI issues, such as who the audience is, what they will find interesting, and what sorts of robot stimulus-action sequences are best suited to both audience interest and intended learning goals, as well as to the crowded, noisy NRW exhibition venue environment. As noted, the NRW audience was primarily children, adolescents, and their parents. Depending on enrollment in any semester, students worked in teams on the final project demonstration. When multiple teams were involved, they coordinated with one another so that their separate projects all could be integrated into one multifaceted exhibit using the DOMER application. The class met twice weekly for an hour and 15 minutes during the first 7 weeks of the semester. During the first weekly meeting, students demonstrated their weekly assignment to the class. In this meeting, students discussed any challenges or questions they may have encountered in completing their weekly tasks. The second meeting of the week was set aside for the students to work on their next assignment and/or their final projects. They explored the various control techniques that would be required, and were encouraged to ask any additional questions that may have come up. In addition to completing their final projects, students were required to submit a final report which described their NRW project, including any relevant documentation or technical details. In addition, the final report needed to describe clearly what resources were used in the project’s creation, what central HRI considerations were addressed, and what lessons were learned along the way. An example of one such lesson from a student team is the following: We found it is a lot easier to do the general and/or bigger movements manually with the animation mode. This is done by: 1. Unstiffening chains in the arms or legs (found in the Motion Widget). 2. Moving the limbs into position. 103


3. 4.

Right-clicking on the Motion Ruler at the top in the Timeline box. Clicking “Whole Body,” “Head,” “Arms,” or “Legs” under “Store joints in keyframe” depending on what parts of the body we had set in the correct position.

Illustrative Course Outcomes Course outcomes are best illustrated by what students learned, as evidenced by their final projects, and by what they said about their class experience. Also relevant were the audience responses to student demonstrations at NRW. Student learning. Final projects contributed a significant amount to a student’s final grade, and as noted, these projects were displayed in a public forum, so there was considerable motivation to ensure these projects were done well. Although students proposed their final projects around mid-semester, the final details of any project may have evolved until completion. Illustrative past NRW demos have included: 

Interactive games, such as Simon Says (see picture of NAO providing a response to a child who holds up a stimulus card at https://drive.google.com/open?id=0By6YGPMc5KzPZ3VzTlhhTFRRNG8). Game development usually required students to take advantage of the robot’s visual and object recognition capabilities, especially when cards were used for visual rather than auditory audience input to the robot, given the noisy exhibition environment.



Various dance moves by the robot (e.g., the “Single Ladies” dance) to accompany sound clips that audience members could select from a menu of choices, either verbally or by using a touch screen interface (see dance only —audio removed—example at https://drive.google.com/open?id=0By6YGPMc5KzPTEZQMTVCeFFuMVU). Development of such dances required sophisticated control of robot movement sequences and coordination with voice and audio.



Popular campus cheers enacted by the robot (see https://drive.google.com/open?id=0By6YGPMc5KzPNmJ1T2VfWncxUFE and https://drive.google.com/open?id=0By6YGPMc5KzPNjJBaU9Oam92d2s). Again, development of these sequences required movement control and coordination with audio and/or voice.



Two robots enacting the University of Notre Dame “alma mater” as is done by students after sporting events (see https://drive.google.com/open?id=0By6YGPMc5KzPYlpLT0c5Mk5MQzA). Development here required movement coordination across robots.



Impersonations where the robot both acted and vocalized like popular characters (e.g., R2D2 from the Star Wars movies; see picture of the end of an impersonation at https://drive.google.com/open?id=0By6YGPMc5KzPbEhWN3ZmRmxnTWs). Impersonations usually were prompted by stimulus cards held up by an audience member. Development here required object recognition and coordination of movement and robot vocalization/audio.

Student final projects reflected student mastery of various capabilities of the NAO platform and design commitments made to specific HRI considerations prompted by the questions noted above. Student reports of their experiences. Students reported positive experiences in this class. Representative student comments across semesters include the following: 

“This class freed me of worry about many of the technical details, since I don’t have that background, and instead let me focus on what I wanted to make NAO do to be informative and entertaining.”



“I really learned a lot about the NAO this semester as indicated by the quality of my demonstration at the NRW and the audience’s reaction to it.” 104




“Being able to work independently with the robots was definitely a learning experience. The weekly assignments at the beginning of the semester were helpful in being able to familiarize myself with the different sensory and motor capabilities of the robot.”



“Thank you so much for giving me the opportunity to take this course and interact with the NAO. It was my first time doing anything related to robotics, so it was a really neat experience, especially in thinking about the interactions between humans and technology.”

From these student comments, one can see that this course did alleviate many of the previously mentioned barriers that otherwise might have prevented these liberal arts students from enjoying a successful experience in a robotics course. Furthermore, many of the students echoed the sentiments captured in the fourth comment above. This practicum course helped students think about HRI as they worked on developing interesting and relevant exhibitions of robot stimulus-action sequences. National Robotics Week feedback. A feedback survey was created and distributed during NRW 2016 to gauge the public’s interest in the Practicum in Robotics booth. The survey contained questions about the attendee’s favorite booths, including rating scales for the Practicum’s booth and their overall NRW experience. All rating scales used a 5-point Likert scale from “very much” (5) to “really not” (1) and included four different categories to be rated: interesting, informative, fun/entertaining, and worth coming. Two open-ended questions assessed what was liked most about their favorite booth as well as the Practicum booth. Responses to these items were coded independently by two researchers using a schema based on the questions. The average Cohen’s Kappa found for the open-ended items was 0.90, a very high level of agreement. In total, 42 attendees filled out the survey at the Practicum booth. For 82% of respondents, it was their first time attending NRW, and 66% of the respondents reported the Practicum booth was their favorite booth. The most cited reason was the interactivity of the robot with the audience (35%). Several parents noted that their children thoroughly enjoyed playing the Simon Says game with the robot. One parent reported, “The kids had a blast interacting in games with the robots.” When asked specifically about the Practicum booth, 41% of the respondents cited they were particularly impressed with the actions that the robot could perform. Further, a few individuals noted that they found the robot to be “cute” or “adorable” and entertaining. Fig. 1 (dark gray bars) shows average ratings of the NRW exhibition as a whole across the four categories. Standard error bars are included in this graph to reflect the individual variations contributing to these average ratings, and indicate a 95% confidence interval. This figure shows that survey respondents were extremely satisfied with the NRW exhibition overall. Over 80% of respondents gave the highest possible rating (5) on all categories except informative. The informative ratings were split between “very informative” (5) and “somewhat informative” (4), with 58% and 36% ratings in those categories, respectively. This figure (light gray bars) also shows the same rating categories for the Practicum booth, which received ratings similar to those of the NRW overall. In this case, more than 85% of the survey respondents assigned the highest rating to all categories except informative. Again, for that category, participants were split between “very informative” (55%) and “somewhat informative” (35%). These high levels of overall satisfaction for both NRW and the Practicum booth seem to suggest a strong public interest both in the NAO robot and its applications, as well as in robotics as a field. The light gray bars also serve to indirectly validate the quality and effectiveness of student learning and effort in the Practicum class. Initiative 2 Conclusions The Practicum in Robotics course appeared to be an engaging and effective way to draw non-engineering, humanities college students into the field of robotics. Similar to the Summer Robotics Camp, the Practicum course addressed several barriers to engagement with robots, including the use of a platform that is both aesthetically pleasing and easy to control. The emphasis to create interest-driven, interactive sequences kept students motivated, as did the fact that their final projects were put to the test in a public forum in front of a live audience. Most students came into the course with an undeveloped interest and understanding of robotics and HRI but left with a broadened

105


Figure 1. Side-by side comparison of the average respondent ratings for the NRW Exhibition overall and the Practicum Booth, polled from the audience at NRW 2016. Error bars indicate a 95% confidence interval. perspective of both. In addition, students interacted with the public and received relevant feedback on their work. All of this added up to a powerful learning experience that many non-technical students reported they greatly valued. By using a visually attractive robot and creating entertaining, interactive sequences that were displayed in a public venue, students in the course provided a valuable service to the local community. Practicum students came to the NRW exhibition not only prepared to demonstrate the NAO robot’s capabilities but also to answer any questions that attendees may have. Attendee questions ranged from how the robots were programmed to what they are used for in practice. Using NAO robots as teaching tools in the exhibition venue helped children and adolescent attendees learn more about robots and HRI, potentially sparking at least some to have further interests in this field. Two noteworthy evolutions have occurred thus far during this course’s tenure. One involved incorporating more self-guided learning as part of the class. In the first few years the class was offered, a more structured approach was used in which specific mini-lectures were employed and exercises or tasks were assigned to each student for each class. During these years, feedback from students indicated a preference for less structure and more independent study. Accordingly, we adjusted the structure by moving to seven specific assignments, using mini-lectures as needed, and allowing one of the two classes each week to be devoted to self-study in a lab-section-like environment. Students indicated that this change has been beneficial both in terms of course flexibility and increased time they have to devote to final projects. 106


A second evolution occurred when we began to link the final project to the NRW exhibition. Prior to that, student projects were created and displayed only for the instructors and other students in the class. There was no public display of projects, nor was there a clear audience toward which the final projects were aimed, other than the instructors. Students have indicated having a visible, public display venue for their final projects has enhanced their motivation to create projects that will both entertain and inform the exhibition audience. For those who might consider a course of this nature on their campuses, we would strongly recommend inclusion of the aforementioned evolutions, if at all possible. Also, we would suggest the following further modifications, which we plan on incorporating into our own course going forward: 

Form learning teams based on experience. Students have come to our class with varying levels of technical experience. We have noted that some liberal arts students have fairly extensive backgrounds in computers and programming, even if not specifically with robotics. When this occurs, such students can be a valuable resource for others in the class. Explicitly creating functional teams by pairing students having more experience with those having less experience is an approach we want to explore going forward when the opportunity presents itself. As noted above, this functional team approach has been used to some extent in robotics competitions, and we believe it may offer significant advantages not only in team competition but also in the learning context of our class.



Devote assigned time to reading and discussion of HRI design issues. While informal discussion of design issues in the context of the HRI considerations and questions posed above has occurred in our class, there is merit to explicit assignment of readings and discussion germane to this topic. We will expand our weekly assignments to include several specific readings and class discussion based on papers such as those by Goodrich and Olsen (2003), Forlizzi et al. (2004), Kahn et al. (2008), Kiesler (2005), and others. We believe these readings and discussions will act as a springboard for more effective design of final projects in the class.

Overall Conclusions Both of the present robotics educational initiatives have attracted students with non-engineering backgrounds who came to these experiences with little or no prior exposure to robotics. Both initiatives afforded students creative license to develop robot applications based on their own unique talents and interests. Similar to Ahlgren and Verner’s (2013) competition, the Summer Robotics Camp and Practicum course shifted the focus of robotics education away from mechanics and technical programming toward the kinds of stimulus-action sequences that would be required to support the desired human-robot interactions with the intended end-user audience. In both initiatives reported here, interest-driven educational methods were used that employed a visual, behavior-based approach to developing interactive control sequences. These strategies helped make the subject of robotics and HRI more accessible to many students who otherwise might be intimidated by the field. As shown by Liu’s (2010) survey, robotics is often viewed as an exclusively engineering topic, which could potentially deter social scientists and others outside of the engineering field from learning about robotics and utilizing such technology to extend their line of research. By drawing more non-traditional students into the field of robotics, HRI applications will become more innovative, thereby further extending the reach of robots in our society. Social robotics is bound to benefit from more interdisciplinary involvement in HRI. As we observed in the Summer Robotics Camp, student pairs were motivated to create various final projects showcasing how robots could interact with the audience in ways demonstrating relevant social or communication skills. Undoubtedly, the camp team members with ASD benefitted from these final projects in multiple ways related to their own social skills and communication competencies. Likewise, in the Practicum course, students created socially entertaining and informative applications for public participants of all ages. In both of these cases, students needed to consider the social goals and needs of their end-user populations, and they received valuable feedback from their respective audiences. Non-traditional robotics students may bring novel perspectives and unique experiences to these kinds of social robotics tasks. Both initiatives reported in this paper combined structured education and self-exploration guided by personal interests. The former method allows for the scaffolding of skills needed to take full advantage of self-exploration. In the Summer Robotics Camp, structure was provided in the first four days by examining specific robot control 107


features and linking them to real-world social or communication skills. In the Practicum course, the first seven weeks were devoted to building basic familiarity with the elements of robot control. Scaffolding in this way allowed relevant skills to build on one another instead of standing alone, which may facilitate student understanding of how various robot control topics fit together. As others have noted, scaffolding allows students to produce highly complex and multidimensional interactive projects (Kandlhofer & Steinbauer, 2016). Scaffolding some initial robotics control skills may allow later self-exploration to be more interesting and productive. In this way, students will have a base of knowledge and experience to fuel their interest-driven exploration. In the Summer Robotics Camp, the structured experiences early in the camp facilitated later work on their interest-driven final projects. Likewise, the initial scaffolding in the Practicum course provided by the weekly assignments set the stage for self-guided work later in the final projects. Consistent with the observations of Cabibihan et al. (2013), we believe this approach allowed students to perform at a higher level in the selfexploratory phase than they would if they had been only assigned a specific project or task. Self-direction in educational pursuits may successfully encourage learning outcomes and products that are unique and innovative. As the field of HRI continues to grow, development methods related to HRI will benefit from broader and more interdisciplinary perspectives in HRI. Drawing non-traditional students into robotics from the social sciences and humanities will greatly contribute to a broadened perspective. The educational programs described in this paper represent ways of accomplishing this goal. Together with the other programs described in this special issue, the field is acquiring the educational tools needed to prepare itself for an exciting, increasingly interdisciplinary, and collaborative future.

Acknowledgements This work was funded by a grant from the Florence V. Carroll Charitable Trust through the Wells Fargo Foundation, the Institute for Scholarship in the Liberal Arts at the University of Notre Dame, the University of Notre Dame Career Center, the Glynn Family Honors program at the University of Notre Dame, and a gift from the Forlenza family. The authors would also like to thank the research assistants and staff, past and present, of both the eMotion and eCognition Lab and the Laboratory for Understanding Neurodevelopment at the University of Notre Dame. Their tireless work ensured the success of initiatives described here as well as with the generation and submission of this document. We also want to thank the children, families, and students who have contributed their time to this research.

References Aldebaran, Inc. (2016). NAOqi Documentation V 2.1. Retrieved from http://doc.aldebaran.com Ahlgren, D. J., & Verner, I. M. (2013). Socially responsible engineering education through assistive robotics projects: The robowaiter competition. International Journal of Social Robotics, 5(1), 127-138. doi:10.1007/s12369-011-0137-4 Bak, J. J., & Siperstein, G. N. (1987). Similarity as a factor effecting change in children's attitudes toward mentally retarded peers. American Journal of Mental Deficiency. 91(5), 524-531. Bauminger-Zviely, N., Eden, S., Zancanaro, M., Weiss, P. L., & Gal, E. (2013). Increasing social engagement in children with high-functioning autism spectrum disorder using collaborative technologies in the school environment. Autism, 17(3), 317-339. doi:10.1177/1362361312472989 Boyd, B. A., Conroy, M. A., Mancil, G. R., Nakao, T., & Alter, P. J. (2007). Effects of circumscribed interests on the social behaviors of children with autism spectrum disorders. Journal of Autism and Developmental Disorders, 37(8), 1550-1561. doi:10.1007/s10803-006-0286-8 Cabibihan, J. J., Javed, H., Ang Jr., M., & Aljunied, S. M. (2013). Why robots? A survey on the roles and benefits of social robots in the therapy of children with autism. International Journal of Social Robotics, 5(4), 593618. doi:10.1007/s12369-013-0202-2 108


Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement 20, 37-46. Crowell, C., Shermerhorn, P., Scheutz, M. & Villano, M. (2009). Social presence effects of gendered voice and robot entities: Perceptions and preconceptions. In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, St. Louis, MO: IEEE-RSJ. Dautenhahn, K. (2003). Roles and functions of robots in human society: Implications from research in autism therapy. Robotica, 21(04), 443-452. doi:10.1017/S0263574703004922 Diehl, J. J., Schmitt, L. M., Villano, M., & Crowell, C. R. (2012). The clinical use of robots for individuals with autism spectrum disorders: A critical review. Research in Autism Spectrum Disorders, 6(1), 249-262. doi:10.1016/j.rasd.2011.05.006 Diehl, J. J., Crowell, C. R., Villano, M., Wier, K., Tang, K., Van Ness, M., . . . Shea, N. (2013, May). Humanoid robots as co-therapists in ABA therapy for children with autism spectrum disorder. International Meeting for Autism Research in San Sebastián, Spain. Dunst, C. J., Trivette, C. M., & Hamby, D. W. (2012). Meta-analysis of studies incorporating the interests of young children with autism spectrum disorders into early intervention practices. Autism Research and Treatment, 2012. doi:10.1155/2012/462531 Eguchi, A. (2015). RoboCupJunior for promoting STEM education, 21st century skills, and technological advancement through robotics competition. Robotics and Autonomous Systems, 75, 692-699. Forlizzi, J., DiSalvo, C., & Gemperle, F. (2004). Assistive robotics and an ecology of elders living independently in their homes. Human-Computer Interaction, 19, 25-59. Goodrich, M. A., & Olsen, D. R. (2003). Seven principles of efficient human robot interaction. In Proceedings of the IEEE International Conference on Systems, Man and Cybernetics, 3943-3948. doi: 10.1109/ICSMC.2003.1244504 Gorostiza, J. F., & Salichs, M. A. (2011). End-user programming of a social robot by dialog. Robotics and Autonomous Systems, 59(12), 1102-1114. doi:10.1016/j.robot.2011.07.009 Gresham, F., & Elliott, S. N. (2008). Social skills improvement system (SSIS) rating scales. Bloomington, MN: Pearson Assessments. Kaboski, J. R., Diehl, J. J., Beriont, J., Crowell, C. R., Villano, M., Wier, K., & Tang, K. (2015). Brief report: A pilot Summer Robotics Camp to reduce social anxiety and improve social/vocational skills in adolescents with ASD. Journal of Autism and Developmental Disorders, 45(12), 3862-3869. Kahn, P. H., Freier, N. G., Kanda, T., Ishiguro, H., Ruckert, J. H., Severson, R. L., & Kane, S. K. (2008). Design patterns for sociality in human-robot interaction. In Proceedings of the 3rd ACM/IEEE Conference on Human-Robot Interaction. 97-104. doi:10.1145/1349822.1349836 Kandlhofer, M., & Steinbauer, G. (2016). Evaluating the impact of educational robotics on pupils’ technical- and social-skills and science related attitudes. Robotics and Autonomous Systems, 75, 679-685. Kiesler, S. (2005). Fostering common ground in human-robot interaction. In Proceedings of the 14th IEEE International Workshop on Robot and Human Interactive Communication (ROMAN). Nashville, TN. Aug 14-15. Kim, E. S., Berkovits, L. D., Bernier, E. P., Leyzberg, D., Shic, F., Paul, R., & Scassellati, B. (2013). Social robots as embedded reinforcers of social behavior in children with autism. Journal of autism and developmental disorders, 43(5), 1038-1049. doi:10.1007/s10803-012-1645-2 Kurth, J. A., & Mastergeorge, A. M. (2010). Academic and cognitive profiles of students with autism: Implications for classroom practice and placement. International Journal of Special Education, 25(2), 8-14. La Greca, A. M., & Stone, W. L. (1993). Social anxiety scale for children-revised: Factor structure and concurrent validity. Journal of Clinical Child Psychology, 22(1), 17-27. doi:10.1207/s15374424jccp2201_2

109


Liu, E. Z. F. (2010). Early adolescents' perceptions of educational robots and learning of robotics. British Journal of Educational Technology, 41(3), E44-E47. doi:10.1111/j.1467-8535.2009.00944.x Magnenat, S., Riedo, F., Bonani, M., & Mondada, F. (2012, May). A programming workshop using the robot “Thymio II”: The effect on the understanding by children. In Proceedings of the IEEE Workshop on Advanced Robotics and Its Social Impacts (ARSO). (pp. 24-29). Munich, Germany. Mori, M. (1970/2005). The uncanny valley. (K. F. MacDorman, & T. Minato, Trans.). Energy, 7, 33–35. Nugent, G., Barker, B., Grandgenett, N., & Welch, G. (2016). Robotics camps, clubs, and competitions: Results from a US robotics project. Robotics and Autonomous Systems, 75, 686-691. Riedo, F., Chevalier, M., Magnenat, S., & Mondada, F. (2013, November). Thymio II, a robot that grows wiser with children. In Proceedings of the 2013 IEEE Workshop on Advanced Robotics and Its Social Impacts. (pp. 187-193). IEEE. doi:10.1109/ARSO.2013.6705527 Tsai, C. C. (2004). Adolescents' perceptions toward the Internet: A 4-T framework. Behavior, 7(4), 458-463.

CyberPsychology

&

University of Notre Dame (2016). Computing & Digital Technologies Minor. Retrieved May 11, 2016 from http://cdt.nd.edu Vandevelde, C., & Saldien, J. (2016, March). An open platform for the design of social robot embodiments for faceto-face communication. In Proceedings of the Eleventh ACM/IEEE International Conference on Human Robot Interaction. (pp. 287-294). IEEE Press. doi:10.1109/HRI.2016.7451764 Villano, M., Crowell, C. R., Wier, K., Tang, K., Thomas, B., Shea, N., . . . Diehl, J. J. (2011, March). DOMER: A Wizard of Oz interface for using interactive robots to scaffold social skills for children with autism spectrum disorders. In Proceedings of the 6th International Conference on Human-Robot Interaction. (pp. 279-280). ACM. doi:10.1145/1957656.1957770 Wainer, J., Ferrari, E., Dautenhahn, K., & Robins, B. (2010). The effectiveness of using a robotics class to foster collaboration among groups of children with autism in an exploratory study. Personal and Ubiquitous Computing, 14(5), 445-455. doi:10.1016/j.rasd.2010.08.002 J. D. Zenk, MA, University of Notre Dame du Lac, [email protected]; C. R. Crowell, PhD, University of Notre Dame du Lac, [email protected]; M. Villano, PhD, University of Notre Dame du Lac, [email protected]; J. Kaboski, PhD, University of Notre Dame du Lac, [email protected]; K. Tang, PhD, University of Notre Dame du Lac, [email protected]; J. Diehl, PhD, University of Notre Dame du Lac & LOGAN Community Resources, Inc., South Bend, IN, [email protected]

110